diff --git a/arc_consistency_heuristics.ipynb b/arc_consistency_heuristics.ipynb
new file mode 100644
index 000000000..fb2241819
--- /dev/null
+++ b/arc_consistency_heuristics.ipynb
@@ -0,0 +1,1999 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "# Constraint Satisfaction Problems\n",
+ "---\n",
+ "# Heuristics for Arc-Consistency Algorithms\n",
+ "\n",
+ "## Introduction\n",
+ "A ***Constraint Satisfaction Problem*** is a triple $(X,D,C)$ where: \n",
+ "- $X$ is a set of variables $X_1, …, X_n$;\n",
+ "- $D$ is a set of domains $D_1, …, D_n$, one for each variable and each of which consists of a set of allowable values $v_1, ..., v_k$;\n",
+ "- $C$ is a set of constraints that specify allowable combinations of values.\n",
+ "\n",
+ "A CSP is called *arc-consistent* if every value in the domain of every variable is supported by all the neighbors of the variable while, is called *inconsistent*, if it has no solutions.
\n",
+ "***Arc-consistency algorithms*** remove all unsupported values from the domains of variables making the CSP *arc-consistent* or decide that a CSP is *inconsistent* by finding that some variable has no supported values in its domain.
\n",
+ "Heuristics significantly enhance the efficiency of the *arc-consistency algorithms* improving their average performance in terms of *consistency-checks* which can be considered a standard measure of goodness for such algorithms. *Arc-heuristic* operate at arc-level and selects the constraint that will be used for the next check, while *domain-heuristics* operate at domain-level and selects which values will be used for the next support-check."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from csp import *"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Domain-Heuristics for Arc-Consistency Algorithms\n",
+ "In [[1]](#cite-van2002domain) are investigated the effects of a *domain-heuristic* based on the notion of a *double-support check* by studying its average time-complexity.\n",
+ "\n",
+ "The objective of *arc-consistency algorithms* is to resolve some uncertainty; it has to be know, for each $v_i \\in D_i$ and for each $v_j \\in D_j$, whether it is supported.\n",
+ "\n",
+ "A *single-support check*, $(v_i, v_j) \\in C_{ij}$, is one in which, before the check is done, it is already known that either $v_i$ or $v_j$ are supported. \n",
+ "\n",
+ "A *double-support check* $(v_i, v_j) \\in C_{ij}$, is one in which there is still, before the check, uncertainty about the support-status of both $v_i$ and $v_j$. \n",
+ "\n",
+ "If a *double-support check* is successful, two uncertainties are resolved. If a *single-support check* is successful, only one uncertainty is resolved. A good *arc-consistency algorithm*, therefore, would always choose to do a *double-support check* in preference of a *single-support check*, because the cormer offers the potential higher payback.\n",
+ "\n",
+ "The improvement with *double-support check* is that, where possible, *consistency-checks* are used to find supports for two values, one value in the domain of each variable, which were previously known to be unsupported. It is motivated by the insight that *in order to minimize the number of consistency-checks it is necessary to maximize the number of uncertainties which are resolved per check*."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "### AC-3b: an improved version of AC-3 with Double-Support Checks"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "As shown in [[2]](#cite-van2000improving) the idea is to use *double-support checks* to improve the average performance of `AC3` which does not exploit the fact that relations are bidirectional and results in a new general purpose *arc-consistency algorithm* called `AC3b`."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mAC3\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdom_j_up\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"[Figure 6.3]\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mqueue\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXk\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvariables\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXk\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mneighbors\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msupport_pruning\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mrevise\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mrevised\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;31m# CSP is inconsistent\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXk\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mneighbors\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mXk\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;31m# CSP is satisfiable\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource AC3"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mrevise\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Return true if we remove a value.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# If Xi=x conflicts with Xj=y for every possible y, eliminate Xi=x\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if all(not csp.constraints(Xi, x, Xj, y) for y in csp.curr_domains[Xj]):\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0my\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconstraints\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mconflict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mconflict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprune\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mrevised\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource revise"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "At any stage in the process of making 2-variable CSP *arc-consistent* in `AC3b`:\n",
+ "- there is a set $S_i^+ \\subseteq D_i$ whose values are all known to be supported by $X_j$;\n",
+ "- there is a set $S_i^? = D_i \\setminus S_i^+$ whose values are unknown, as yet, to be supported by $X_j$.\n",
+ "\n",
+ "The same holds if the roles for $X_i$ and $X_j$ are exchanged.\n",
+ "\n",
+ "In order to establish support for a value $v_i^? \\in S_i^?$ it seems better to try to find a support among the values in $S_j^?$ first, because for each $v_j^? \\in S_j^?$ the check $(v_i^?,v_j^?) \\in C_{ij}$ is a *double-support check* and it is just as likely that any $v_j^? \\in S_j^?$ supports $v_i^?$ than it is that any $v_j^+ \\in S_j^+$ does. Only if no support can be found among the elements in $S_j^?$, should the elements $v_j^+$ in $S_j^+$ be used for *single-support checks* $(v_i^?,v_j^+) \\in C_{ij}$. After it has been decided for each value in $D_i$ whether it is supported or not, either $S_x^+ = \\emptyset$ and the 2-variable CSP is *inconsistent*, or $S_x^+ \\neq \\emptyset$ and the CSP is *satisfiable*. In the latter case, the elements from $D_i$ which are supported by $j$ are given by $S_x^+$. The elements in $D_j$ which are supported by $x$ are given by the union of $S_j^+$ with the set of those elements of $S_j^?$ which further processing will show to be supported by some $v_i^+ \\in S_x^+$."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mAC3b\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdom_j_up\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mqueue\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXk\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvariables\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXk\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mneighbors\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msupport_pruning\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Si_p values are all known to be supported by Xj\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Sj_p values are all known to be supported by Xi\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Dj - Sj_p = Sj_u values are unknown, as yet, to be supported by Xi\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mSi_p\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSj_p\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSj_u\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpartition\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mSi_p\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;31m# CSP is inconsistent\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mSi_p\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprune\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mrevised\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXk\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mneighbors\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mXk\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mqueue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# or queue -= {(Xj, Xi)} or queue.remove((Xj, Xi))\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdifference_update\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdifference_update\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# the elements in D_j which are supported by Xi are given by the union of Sj_p with the set of those\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# elements of Sj_u which further processing will show to be supported by some vi_p in Si_p\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvj_p\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mSj_u\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvi_p\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mSi_p\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconstraints\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvj_p\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvi_p\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mSj_p\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvj_p\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mconflict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mSj_p\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprune\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mrevised\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXk\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mneighbors\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mXk\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXk\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;31m# CSP is satisfiable\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource AC3b"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mpartition\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mSi_p\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mSj_p\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mSj_u\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvi_u\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# now, in order to establish support for a value vi_u in Di it seems better to try to find a support among\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# the values in Sj_u first, because for each vj_u in Sj_u the check (vi_u, vj_u) is a double-support check\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# and it is just as likely that any vj_u in Sj_u supports vi_u than it is that any vj_p in Sj_p does...\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvj_u\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mSj_u\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mSj_p\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# double-support check\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconstraints\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvi_u\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvj_u\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mSi_p\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvi_u\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mSj_p\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvj_u\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mconflict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# ... and only if no support can be found among the elements in Sj_u, should the elements vj_p in Sj_p be used\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# for single-support checks (vi_u, vj_p)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mconflict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvj_p\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mSj_p\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# single-support check\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconstraints\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvi_u\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvj_p\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mSi_p\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvi_u\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mconflict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mSi_p\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSj_p\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSj_u\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mSj_p\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource partition"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "`AC3b` is a refinement of the `AC3` algorithm which consists of the fact that if, when arc $(i,j)$ is being processed and the reverse arc $(j,i)$ is also in the queue, then consistency-checks can be saved because only support for the elements in $S_j^?$ has to be found (as opposed to support for all the elements in $D_j$ in the\n",
+ "`AC3` algorithm).
\n",
+ "`AC3b` inherits all its properties like $\\mathcal{O}(ed^3)$ time-complexity and $\\mathcal{O}(e + nd)$ space-complexity fron `AC3` and where $n$ denotes the number of variables in the CSP, $e$ denotes the number of binary constraints and $d$ denotes the maximum domain-size of the variables."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "## Arc-Heuristics for Arc-Consistency Algorithms"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "Many *arc-heuristics* can be devised, based on three major features of CSPs:\n",
+ "- the number of acceptable pairs in each constraint (the *constraint size* or *satisfiability*);\n",
+ "- the *domain size*;\n",
+ "- the number of binary constraints that each variable participates in, equal to the *degree* of the node of that variable in the constraint graph. \n",
+ "\n",
+ "Simple examples of heuristics that might be expected to improve the efficiency of relaxation are:\n",
+ "- ordering the list of variable pairs by *increasing* relative *satisfiability*;\n",
+ "- ordering by *increasing size of the domain* of the variable $v_j$ relaxed against $v_i$;\n",
+ "- ordering by *descending degree* of node of the variable relaxed.\n",
+ "\n",
+ "In [[3]](#cite-wallace1992ordering) are investigated the effects of these *arc-heuristics* in an empirical way, experimenting the effects of them on random CSPs. Their results demonstrate that the first two, later called `sat up` and `dom j up` for n-ary and binary CSPs respectively, significantly reduce the number of *consistency-checks*."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mdom_j_up\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mSortedSet\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mqueue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mneg\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mt\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource dom_j_up"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0msat_up\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mto_do\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mSortedSet\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mto_do\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mvar\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvar\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mscope\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource sat_up"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "## Experimental Results"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "For the experiments below on binary CSPs, in addition to the two *arc-consistency algorithms* already cited above, `AC3` and `AC3b`, the `AC4` algorithm was used.
\n",
+ "The `AC4` algorithm runs in $\\mathcal{O}(ed^2)$ worst-case time but can be slower than `AC3` on average cases."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mAC4\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdom_j_up\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mqueue\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXk\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvariables\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXk\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mneighbors\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msupport_pruning\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msupport_counter\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mvariable_value_pairs_supported\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdefaultdict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0munsupported_variable_value_pairs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# construction and initialization of support sets\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mqueue\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0my\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconstraints\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msupport_counter\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mvariable_value_pairs_supported\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0msupport_counter\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprune\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0munsupported_variable_value_pairs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mrevised\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;31m# CSP is inconsistent\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# propagation of removed values\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0munsupported_variable_value_pairs\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0munsupported_variable_value_pairs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mvariable_value_pairs_supported\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msupport_counter\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m-=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0msupport_counter\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mXj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprune\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremovals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrevised\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0munsupported_variable_value_pairs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mrevised\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurr_domains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mXi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;31m# CSP is inconsistent\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;31m# CSP is satisfiable\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource AC4"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Sudoku"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "#### Easy Sudoku"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ ". . 3 | . 2 . | 6 . .\n",
+ "9 . . | 3 . 5 | . . 1\n",
+ ". . 1 | 8 . 6 | 4 . .\n",
+ "------+-------+------\n",
+ ". . 8 | 1 . 2 | 9 . .\n",
+ "7 . . | . . . | . . 8\n",
+ ". . 6 | 7 . 8 | 2 . .\n",
+ "------+-------+------\n",
+ ". . 2 | 6 . 9 | 5 . .\n",
+ "8 . . | 2 . 3 | . . 9\n",
+ ". . 5 | . 1 . | 3 . .\n"
+ ]
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(easy1)\n",
+ "sudoku.display(sudoku.infer_assignment())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 23.6 ms, sys: 0 ns, total: 23.6 ms\n",
+ "Wall time: 22.4 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3 needs 11322 consistency-checks'"
+ ]
+ },
+ "execution_count": 10,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, checks = AC3(sudoku, arc_heuristic=no_arc_heuristic)\n",
+ "f'AC3 needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 7.43 ms, sys: 3.68 ms, total: 11.1 ms\n",
+ "Wall time: 10.7 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b needs 8345 consistency-checks'"
+ ]
+ },
+ "execution_count": 11,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(easy1)\n",
+ "%time _, checks = AC3b(sudoku, arc_heuristic=no_arc_heuristic)\n",
+ "f'AC3b needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 56.3 ms, sys: 0 ns, total: 56.3 ms\n",
+ "Wall time: 55.4 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC4 needs 27718 consistency-checks'"
+ ]
+ },
+ "execution_count": 12,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(easy1)\n",
+ "%time _, checks = AC4(sudoku, arc_heuristic=no_arc_heuristic)\n",
+ "f'AC4 needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 13,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 17.2 ms, sys: 0 ns, total: 17.2 ms\n",
+ "Wall time: 16.9 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3 with DOM J UP arc heuristic needs 6925 consistency-checks'"
+ ]
+ },
+ "execution_count": 13,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(easy1)\n",
+ "%time _, checks = AC3(sudoku, arc_heuristic=dom_j_up)\n",
+ "f'AC3 with DOM J UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 40.9 ms, sys: 2.47 ms, total: 43.4 ms\n",
+ "Wall time: 41.7 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b with DOM J UP arc heuristic needs 6278 consistency-checks'"
+ ]
+ },
+ "execution_count": 14,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(easy1)\n",
+ "%time _, checks = AC3b(sudoku, arc_heuristic=dom_j_up)\n",
+ "f'AC3b with DOM J UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 15,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 38.9 ms, sys: 1.96 ms, total: 40.9 ms\n",
+ "Wall time: 40.7 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC4 with DOM J UP arc heuristic needs 9393 consistency-checks'"
+ ]
+ },
+ "execution_count": 15,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(easy1)\n",
+ "%time _, checks = AC4(sudoku, arc_heuristic=dom_j_up)\n",
+ "f'AC4 with DOM J UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 16,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "4 8 3 | 9 2 1 | 6 5 7\n",
+ "9 6 7 | 3 4 5 | 8 2 1\n",
+ "2 5 1 | 8 7 6 | 4 9 3\n",
+ "------+-------+------\n",
+ "5 4 8 | 1 3 2 | 9 7 6\n",
+ "7 2 9 | 5 6 4 | 1 3 8\n",
+ "1 3 6 | 7 9 8 | 2 4 5\n",
+ "------+-------+------\n",
+ "3 7 2 | 6 8 9 | 5 1 4\n",
+ "8 1 4 | 2 5 3 | 7 6 9\n",
+ "6 9 5 | 4 1 7 | 3 8 2\n"
+ ]
+ }
+ ],
+ "source": [
+ "backtracking_search(sudoku, select_unassigned_variable=mrv, inference=forward_checking)\n",
+ "sudoku.display(sudoku.infer_assignment())"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "#### Harder Sudoku"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 17,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "4 1 7 | 3 6 9 | 8 . 5\n",
+ ". 3 . | . . . | . . .\n",
+ ". . . | 7 . . | . . .\n",
+ "------+-------+------\n",
+ ". 2 . | . . . | . 6 .\n",
+ ". . . | . 8 . | 4 . .\n",
+ ". . . | . 1 . | . . .\n",
+ "------+-------+------\n",
+ ". . . | 6 . 3 | . 7 .\n",
+ "5 . . | 2 . . | . . .\n",
+ "1 . 4 | . . . | . . .\n"
+ ]
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(harder1)\n",
+ "sudoku.display(sudoku.infer_assignment())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 18,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 17.7 ms, sys: 481 µs, total: 18.2 ms\n",
+ "Wall time: 17.2 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3 needs 12837 consistency-checks'"
+ ]
+ },
+ "execution_count": 18,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, checks = AC3(sudoku, arc_heuristic=no_arc_heuristic)\n",
+ "f'AC3 needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 24.1 ms, sys: 2.6 ms, total: 26.7 ms\n",
+ "Wall time: 25.1 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b needs 8864 consistency-checks'"
+ ]
+ },
+ "execution_count": 19,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(harder1)\n",
+ "%time _, checks = AC3b(sudoku, arc_heuristic=no_arc_heuristic)\n",
+ "f'AC3b needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 63.4 ms, sys: 3.48 ms, total: 66.9 ms\n",
+ "Wall time: 65.5 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC4 needs 44213 consistency-checks'"
+ ]
+ },
+ "execution_count": 20,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(harder1)\n",
+ "%time _, checks = AC4(sudoku, arc_heuristic=no_arc_heuristic)\n",
+ "f'AC4 needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 21,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 9.96 ms, sys: 570 µs, total: 10.5 ms\n",
+ "Wall time: 10.3 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3 with DOM J UP arc heuristic needs 7045 consistency-checks'"
+ ]
+ },
+ "execution_count": 21,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(harder1)\n",
+ "%time _, checks = AC3(sudoku, arc_heuristic=dom_j_up)\n",
+ "f'AC3 with DOM J UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 36.1 ms, sys: 0 ns, total: 36.1 ms\n",
+ "Wall time: 35.5 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b with DOM J UP arc heuristic needs 6994 consistency-checks'"
+ ]
+ },
+ "execution_count": 22,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(harder1)\n",
+ "%time _, checks = AC3b(sudoku, arc_heuristic=dom_j_up)\n",
+ "f'AC3b with DOM J UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 40.3 ms, sys: 0 ns, total: 40.3 ms\n",
+ "Wall time: 39.7 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC4 with DOM J UP arc heuristic needs 19210 consistency-checks'"
+ ]
+ },
+ "execution_count": 23,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "sudoku = Sudoku(harder1)\n",
+ "%time _, checks = AC4(sudoku, arc_heuristic=dom_j_up)\n",
+ "f'AC4 with DOM J UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "4 1 7 | 3 6 9 | 8 2 5\n",
+ "6 3 2 | 1 5 8 | 9 4 7\n",
+ "9 5 8 | 7 2 4 | 3 1 6\n",
+ "------+-------+------\n",
+ "8 2 5 | 4 3 7 | 1 6 9\n",
+ "7 9 1 | 5 8 6 | 4 3 2\n",
+ "3 4 6 | 9 1 2 | 7 5 8\n",
+ "------+-------+------\n",
+ "2 8 9 | 6 4 3 | 5 7 1\n",
+ "5 7 3 | 2 9 1 | 6 8 4\n",
+ "1 6 4 | 8 7 5 | 2 9 3\n"
+ ]
+ }
+ ],
+ "source": [
+ "backtracking_search(sudoku, select_unassigned_variable=mrv, inference=forward_checking)\n",
+ "sudoku.display(sudoku.infer_assignment())"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "### 8 Queens"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 27,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ ". - . - . - . - 0 0 0 0 0 0 0 0 \n",
+ "- . - . - . - . 0 0 0 0 0 0 0 0 \n",
+ ". - . - . - . - 0 0 0 0 0 0 0 0 \n",
+ "- . - . - . - . 0 0 0 0 0 0 0 0 \n",
+ ". - . - . - . - 0 0 0 0 0 0 0 0 \n",
+ "- . - . - . - . 0 0 0 0 0 0 0 0 \n",
+ ". - . - . - . - 0 0 0 0 0 0 0 0 \n",
+ "- . - . - . - . 0 0 0 0 0 0 0 0 \n"
+ ]
+ }
+ ],
+ "source": [
+ "chess = NQueensCSP(8)\n",
+ "chess.display(chess.infer_assignment())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 689 µs, sys: 193 µs, total: 882 µs\n",
+ "Wall time: 892 µs\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3 needs 666 consistency-checks'"
+ ]
+ },
+ "execution_count": 28,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, checks = AC3(chess, arc_heuristic=no_arc_heuristic)\n",
+ "f'AC3 needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 30,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 451 µs, sys: 127 µs, total: 578 µs\n",
+ "Wall time: 584 µs\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b needs 428 consistency-checks'"
+ ]
+ },
+ "execution_count": 30,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "chess = NQueensCSP(8)\n",
+ "%time _, checks = AC3b(chess, arc_heuristic=no_arc_heuristic)\n",
+ "f'AC3b needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 32,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 8.53 ms, sys: 109 µs, total: 8.64 ms\n",
+ "Wall time: 8.48 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC4 needs 4096 consistency-checks'"
+ ]
+ },
+ "execution_count": 32,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "chess = NQueensCSP(8)\n",
+ "%time _, checks = AC4(chess, arc_heuristic=no_arc_heuristic)\n",
+ "f'AC4 needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 34,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 1.88 ms, sys: 0 ns, total: 1.88 ms\n",
+ "Wall time: 1.88 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3 with DOM J UP arc heuristic needs 666 consistency-checks'"
+ ]
+ },
+ "execution_count": 34,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "chess = NQueensCSP(8)\n",
+ "%time _, checks = AC3(chess, arc_heuristic=dom_j_up)\n",
+ "f'AC3 with DOM J UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 36,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 1.21 ms, sys: 326 µs, total: 1.53 ms\n",
+ "Wall time: 1.54 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b with DOM J UP arc heuristic needs 792 consistency-checks'"
+ ]
+ },
+ "execution_count": 36,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "chess = NQueensCSP(8)\n",
+ "%time _, checks = AC3b(chess, arc_heuristic=dom_j_up)\n",
+ "f'AC3b with DOM J UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 38,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 4.71 ms, sys: 0 ns, total: 4.71 ms\n",
+ "Wall time: 4.65 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC4 with DOM J UP arc heuristic needs 4096 consistency-checks'"
+ ]
+ },
+ "execution_count": 38,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "chess = NQueensCSP(8)\n",
+ "%time _, checks = AC4(chess, arc_heuristic=dom_j_up)\n",
+ "f'AC4 with DOM J UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 39,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ ". - . - Q - . - 2 2 3 3 0* 1 1 2 \n",
+ "- Q - . - . - . 1 0* 3 3 2 2 2 2 \n",
+ ". - . - . Q . - 3 2 3 2 2 0* 3 2 \n",
+ "Q . - . - . - . 0* 3 1 2 3 3 3 3 \n",
+ ". - . - . - Q - 2 2 2 2 3 3 0* 2 \n",
+ "- . - Q - . - . 2 1 3 0* 2 3 2 2 \n",
+ ". - . - . - . Q 1 3 2 3 3 1 2 0* \n",
+ "- . Q . - . - . 2 2 0* 2 2 2 2 2 \n"
+ ]
+ }
+ ],
+ "source": [
+ "backtracking_search(chess, select_unassigned_variable=mrv, inference=forward_checking)\n",
+ "chess.display(chess.infer_assignment())"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "For the experiments below on n-ary CSPs, due to the n-ary constraints, the `GAC` algorithm was used.
\n",
+ "The `GAC` algorithm has $\\mathcal{O}(er^2d^t)$ time-complexity and $\\mathcal{O}(erd)$ space-complexity where $e$ denotes the number of n-ary constraints, $r$ denotes the constraint arity and $d$ denotes the maximum domain-size of the variables."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 40,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ " \u001b[0;32mdef\u001b[0m \u001b[0mGAC\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0morig_domains\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mto_do\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msat_up\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Makes this CSP arc-consistent using Generalized Arc Consistency\u001b[0m\n",
+ "\u001b[0;34m orig_domains is the original domains\u001b[0m\n",
+ "\u001b[0;34m to_do is a set of (variable,constraint) pairs\u001b[0m\n",
+ "\u001b[0;34m returns the reduced domains (an arc-consistent variable:domain dictionary)\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0morig_domains\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0morig_domains\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdomains\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mto_do\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mto_do\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mconst\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mconst\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconstraints\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvar\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mconst\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mscope\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mto_do\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mto_do\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomains\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0morig_domains\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mto_do\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mto_do\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0mto_do\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mconst\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mto_do\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mother_vars\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mov\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mov\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mconst\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mscope\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mov\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mvar\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_domain\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mother_vars\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mval\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdomains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mconst\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mholds\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mval\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_domain\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# new_domain = {val for val in domains[var]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if const.holds({var: val})}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mother_vars\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mother\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mother_vars\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mval\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdomains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mother_val\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdomains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mother\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mconst\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mholds\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mother\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mother_val\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_domain\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# new_domain = {val for val in domains[var]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if any(const.holds({var: val, other: other_val})\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# for other_val in domains[other])}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;31m# general case\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mval\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdomains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mholds\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0many_holds\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdomains\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mconst\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mval\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mother_vars\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mchecks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mholds\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_domain\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# new_domain = {val for val in domains[var]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if self.any_holds(domains, const, {var: val}, other_vars)}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mnew_domain\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mdomains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomains\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnew_domain\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mnew_domain\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdomains\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0madd_to_do\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnew_to_do\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mconst\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdifference\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mto_do\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mto_do\u001b[0m \u001b[0;34m|=\u001b[0m \u001b[0madd_to_do\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdomains\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mchecks\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource ACSolver.GAC"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "### Crossword"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 41,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[_] [_] [_] [*] [*] \n",
+ "[_] [*] [_] [*] [*] \n",
+ "[_] [_] [_] [_] [*] \n",
+ "[_] [*] [_] [*] [*] \n",
+ "[*] [*] [_] [_] [_] \n",
+ "[*] [*] [_] [*] [*] \n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "{'ant',\n",
+ " 'big',\n",
+ " 'book',\n",
+ " 'bus',\n",
+ " 'buys',\n",
+ " 'car',\n",
+ " 'ginger',\n",
+ " 'has',\n",
+ " 'hold',\n",
+ " 'lane',\n",
+ " 'search',\n",
+ " 'symbol',\n",
+ " 'syntax',\n",
+ " 'year'}"
+ ]
+ },
+ "execution_count": 41,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "crossword = Crossword(crossword1, words1)\n",
+ "crossword.display()\n",
+ "words1"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 36,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 1min 20s, sys: 2.02 ms, total: 1min 20s\n",
+ "Wall time: 1min 20s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC needs 64617645 consistency-checks'"
+ ]
+ },
+ "execution_count": 36,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, _, checks = ACSolver(crossword).GAC(arc_heuristic=no_heuristic)\n",
+ "f'GAC needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 42,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 1.19 s, sys: 0 ns, total: 1.19 s\n",
+ "Wall time: 1.19 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC with SAT UP arc heuristic needs 908015 consistency-checks'"
+ ]
+ },
+ "execution_count": 42,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "crossword = Crossword(crossword1, words1)\n",
+ "%time _, _, checks = ACSolver(crossword).GAC(arc_heuristic=sat_up)\n",
+ "f'GAC with SAT UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 43,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[B] [U] [S] [*] [*] \n",
+ "[U] [*] [E] [*] [*] \n",
+ "[Y] [E] [A] [R] [*] \n",
+ "[S] [*] [R] [*] [*] \n",
+ "[*] [*] [C] [A] [R] \n",
+ "[*] [*] [H] [*] [*] \n"
+ ]
+ }
+ ],
+ "source": [
+ "crossword.display(ACSolver(crossword).domain_splitting())"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "### Kakuro"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### Easy Kakuro"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 44,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[*]\t10\\\t13\\\t[*]\t\n",
+ "\\3\t[_]\t[_]\t13\\\t\n",
+ "\\12\t[_]\t[_]\t[_]\t\n",
+ "\\21\t[_]\t[_]\t[_]\t\n"
+ ]
+ }
+ ],
+ "source": [
+ "kakuro = Kakuro(kakuro2)\n",
+ "kakuro.display()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 45,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 17.8 ms, sys: 171 µs, total: 18 ms\n",
+ "Wall time: 16.4 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC needs 2752 consistency-checks'"
+ ]
+ },
+ "execution_count": 45,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, _, checks = ACSolver(kakuro).GAC(arc_heuristic=no_heuristic)\n",
+ "f'GAC needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 46,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 8.55 ms, sys: 0 ns, total: 8.55 ms\n",
+ "Wall time: 8.39 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC with SAT UP arc heuristic needs 1765 consistency-checks'"
+ ]
+ },
+ "execution_count": 46,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "kakuro = Kakuro(kakuro2)\n",
+ "%time _, _, checks = ACSolver(kakuro).GAC(arc_heuristic=sat_up)\n",
+ "f'GAC with SAT UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 47,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[*]\t10\\\t13\\\t[*]\t\n",
+ "\\3\t[1]\t[2]\t13\\\t\n",
+ "\\12\t[5]\t[3]\t[4]\t\n",
+ "\\21\t[4]\t[8]\t[9]\t\n"
+ ]
+ }
+ ],
+ "source": [
+ "kakuro.display(ACSolver(kakuro).domain_splitting())"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "#### Medium Kakuro"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 48,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[*]\t17\\\t28\\\t[*]\t42\\\t22\\\t\n",
+ "\\9\t[_]\t[_]\t31\\14\t[_]\t[_]\t\n",
+ "\\20\t[_]\t[_]\t[_]\t[_]\t[_]\t\n",
+ "[*]\t\\30\t[_]\t[_]\t[_]\t[_]\t\n",
+ "[*]\t22\\24\t[_]\t[_]\t[_]\t[*]\t\n",
+ "\\25\t[_]\t[_]\t[_]\t[_]\t11\\\t\n",
+ "\\20\t[_]\t[_]\t[_]\t[_]\t[_]\t\n",
+ "\\14\t[_]\t[_]\t\\17\t[_]\t[_]\t\n"
+ ]
+ }
+ ],
+ "source": [
+ "kakuro = Kakuro(kakuro3)\n",
+ "kakuro.display()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 49,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 1.96 s, sys: 0 ns, total: 1.96 s\n",
+ "Wall time: 1.96 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC needs 1290179 consistency-checks'"
+ ]
+ },
+ "execution_count": 49,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, _, checks = ACSolver(kakuro).GAC(arc_heuristic=no_heuristic)\n",
+ "f'GAC needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 50,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 225 ms, sys: 0 ns, total: 225 ms\n",
+ "Wall time: 223 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC with SAT UP arc heuristic needs 148780 consistency-checks'"
+ ]
+ },
+ "execution_count": 50,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "kakuro = Kakuro(kakuro3)\n",
+ "%time _, _, checks = ACSolver(kakuro).GAC(arc_heuristic=sat_up)\n",
+ "f'GAC with SAT UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 51,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[*]\t17\\\t28\\\t[*]\t42\\\t22\\\t\n",
+ "\\9\t[8]\t[1]\t31\\14\t[5]\t[9]\t\n",
+ "\\20\t[9]\t[2]\t[1]\t[3]\t[5]\t\n",
+ "[*]\t\\30\t[6]\t[9]\t[7]\t[8]\t\n",
+ "[*]\t22\\24\t[7]\t[8]\t[9]\t[*]\t\n",
+ "\\25\t[8]\t[4]\t[7]\t[6]\t11\\\t\n",
+ "\\20\t[5]\t[3]\t[6]\t[4]\t[2]\t\n",
+ "\\14\t[9]\t[5]\t\\17\t[8]\t[9]\t\n"
+ ]
+ }
+ ],
+ "source": [
+ "kakuro.display(ACSolver(kakuro).domain_splitting())"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "#### Harder Kakuro"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 52,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[*]\t[*]\t[*]\t[*]\t[*]\t4\\\t24\\\t11\\\t[*]\t[*]\t[*]\t11\\\t17\\\t[*]\t[*]\t\n",
+ "[*]\t[*]\t[*]\t17\\\t11\\12\t[_]\t[_]\t[_]\t[*]\t[*]\t24\\10\t[_]\t[_]\t11\\\t[*]\t\n",
+ "[*]\t4\\\t16\\26\t[_]\t[_]\t[_]\t[_]\t[_]\t[*]\t\\20\t[_]\t[_]\t[_]\t[_]\t16\\\t\n",
+ "\\20\t[_]\t[_]\t[_]\t[_]\t24\\13\t[_]\t[_]\t16\\\t\\12\t[_]\t[_]\t23\\10\t[_]\t[_]\t\n",
+ "\\10\t[_]\t[_]\t24\\12\t[_]\t[_]\t16\\5\t[_]\t[_]\t16\\30\t[_]\t[_]\t[_]\t[_]\t[_]\t\n",
+ "[*]\t[*]\t3\\26\t[_]\t[_]\t[_]\t[_]\t\\12\t[_]\t[_]\t4\\\t16\\14\t[_]\t[_]\t[*]\t\n",
+ "[*]\t\\8\t[_]\t[_]\t\\15\t[_]\t[_]\t34\\26\t[_]\t[_]\t[_]\t[_]\t[_]\t[*]\t[*]\t\n",
+ "[*]\t\\11\t[_]\t[_]\t3\\\t17\\\t\\14\t[_]\t[_]\t\\8\t[_]\t[_]\t7\\\t17\\\t[*]\t\n",
+ "[*]\t[*]\t[*]\t23\\10\t[_]\t[_]\t3\\9\t[_]\t[_]\t4\\\t23\\\t\\13\t[_]\t[_]\t[*]\t\n",
+ "[*]\t[*]\t10\\26\t[_]\t[_]\t[_]\t[_]\t[_]\t\\7\t[_]\t[_]\t30\\9\t[_]\t[_]\t[*]\t\n",
+ "[*]\t17\\11\t[_]\t[_]\t11\\\t24\\8\t[_]\t[_]\t11\\21\t[_]\t[_]\t[_]\t[_]\t16\\\t17\\\t\n",
+ "\\29\t[_]\t[_]\t[_]\t[_]\t[_]\t\\7\t[_]\t[_]\t23\\14\t[_]\t[_]\t3\\17\t[_]\t[_]\t\n",
+ "\\10\t[_]\t[_]\t3\\10\t[_]\t[_]\t[*]\t\\8\t[_]\t[_]\t4\\25\t[_]\t[_]\t[_]\t[_]\t\n",
+ "[*]\t\\16\t[_]\t[_]\t[_]\t[_]\t[*]\t\\23\t[_]\t[_]\t[_]\t[_]\t[_]\t[*]\t[*]\t\n",
+ "[*]\t[*]\t\\6\t[_]\t[_]\t[*]\t[*]\t\\15\t[_]\t[_]\t[_]\t[*]\t[*]\t[*]\t[*]\t\n"
+ ]
+ }
+ ],
+ "source": [
+ "kakuro = Kakuro(kakuro4)\n",
+ "kakuro.display()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 53,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 76.5 ms, sys: 847 µs, total: 77.4 ms\n",
+ "Wall time: 77 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC needs 46633 consistency-checks'"
+ ]
+ },
+ "execution_count": 53,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, _, checks = ACSolver(kakuro).GAC()\n",
+ "f'GAC needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 54,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 64.6 ms, sys: 0 ns, total: 64.6 ms\n",
+ "Wall time: 63.6 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC with SAT UP arc heuristic needs 36828 consistency-checks'"
+ ]
+ },
+ "execution_count": 54,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "kakuro = Kakuro(kakuro4)\n",
+ "%time _, _, checks = ACSolver(kakuro).GAC(arc_heuristic=sat_up)\n",
+ "f'GAC with SAT UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 55,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "[*]\t[*]\t[*]\t[*]\t[*]\t4\\\t24\\\t11\\\t[*]\t[*]\t[*]\t11\\\t17\\\t[*]\t[*]\t\n",
+ "[*]\t[*]\t[*]\t17\\\t11\\12\t[3]\t[7]\t[2]\t[*]\t[*]\t24\\10\t[2]\t[8]\t11\\\t[*]\t\n",
+ "[*]\t4\\\t16\\26\t[8]\t[5]\t[1]\t[9]\t[3]\t[*]\t\\20\t[8]\t[1]\t[9]\t[2]\t16\\\t\n",
+ "\\20\t[3]\t[7]\t[9]\t[1]\t24\\13\t[8]\t[5]\t16\\\t\\12\t[9]\t[3]\t23\\10\t[3]\t[7]\t\n",
+ "\\10\t[1]\t[9]\t24\\12\t[3]\t[9]\t16\\5\t[1]\t[4]\t16\\30\t[7]\t[5]\t[8]\t[1]\t[9]\t\n",
+ "[*]\t[*]\t3\\26\t[8]\t[2]\t[7]\t[9]\t\\12\t[3]\t[9]\t4\\\t16\\14\t[9]\t[5]\t[*]\t\n",
+ "[*]\t\\8\t[1]\t[7]\t\\15\t[8]\t[7]\t34\\26\t[1]\t[7]\t[3]\t[9]\t[6]\t[*]\t[*]\t\n",
+ "[*]\t\\11\t[2]\t[9]\t3\\\t17\\\t\\14\t[8]\t[6]\t\\8\t[1]\t[7]\t7\\\t17\\\t[*]\t\n",
+ "[*]\t[*]\t[*]\t23\\10\t[1]\t[9]\t3\\9\t[7]\t[2]\t4\\\t23\\\t\\13\t[4]\t[9]\t[*]\t\n",
+ "[*]\t[*]\t10\\26\t[6]\t[2]\t[8]\t[1]\t[9]\t\\7\t[1]\t[6]\t30\\9\t[1]\t[8]\t[*]\t\n",
+ "[*]\t17\\11\t[3]\t[8]\t11\\\t24\\8\t[2]\t[6]\t11\\21\t[3]\t[9]\t[7]\t[2]\t16\\\t17\\\t\n",
+ "\\29\t[8]\t[2]\t[9]\t[3]\t[7]\t\\7\t[4]\t[3]\t23\\14\t[8]\t[6]\t3\\17\t[9]\t[8]\t\n",
+ "\\10\t[9]\t[1]\t3\\10\t[2]\t[8]\t[*]\t\\8\t[2]\t[6]\t4\\25\t[8]\t[1]\t[7]\t[9]\t\n",
+ "[*]\t\\16\t[4]\t[2]\t[1]\t[9]\t[*]\t\\23\t[1]\t[8]\t[3]\t[9]\t[2]\t[*]\t[*]\t\n",
+ "[*]\t[*]\t\\6\t[1]\t[5]\t[*]\t[*]\t\\15\t[5]\t[9]\t[1]\t[*]\t[*]\t[*]\t[*]\t\n"
+ ]
+ }
+ ],
+ "source": [
+ "kakuro.display(ACSolver(kakuro).domain_splitting())"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "### Cryptarithmetic Puzzle"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "$$\n",
+ "\\begin{array}{@{}r@{}}\n",
+ " S E N D \\\\\n",
+ "{} + M O R E \\\\\n",
+ " \\hline\n",
+ " M O N E Y\n",
+ "\\end{array}\n",
+ "$$"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 57,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [],
+ "source": [
+ "cryptarithmetic = NaryCSP(\n",
+ " {'S': set(range(1, 10)), 'M': set(range(1, 10)),\n",
+ " 'E': set(range(0, 10)), 'N': set(range(0, 10)), 'D': set(range(0, 10)),\n",
+ " 'O': set(range(0, 10)), 'R': set(range(0, 10)), 'Y': set(range(0, 10)),\n",
+ " 'C1': set(range(0, 2)), 'C2': set(range(0, 2)), 'C3': set(range(0, 2)),\n",
+ " 'C4': set(range(0, 2))},\n",
+ " [Constraint(('S', 'E', 'N', 'D', 'M', 'O', 'R', 'Y'), all_diff),\n",
+ " Constraint(('D', 'E', 'Y', 'C1'), lambda d, e, y, c1: d + e == y + 10 * c1),\n",
+ " Constraint(('N', 'R', 'E', 'C1', 'C2'), lambda n, r, e, c1, c2: c1 + n + r == e + 10 * c2),\n",
+ " Constraint(('E', 'O', 'N', 'C2', 'C3'), lambda e, o, n, c2, c3: c2 + e + o == n + 10 * c3),\n",
+ " Constraint(('S', 'M', 'O', 'C3', 'C4'), lambda s, m, o, c3, c4: c3 + s + m == o + 10 * c4),\n",
+ " Constraint(('M', 'C4'), eq)])"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 52,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 21.7 s, sys: 0 ns, total: 21.7 s\n",
+ "Wall time: 21.7 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC needs 14080592 consistency-checks'"
+ ]
+ },
+ "execution_count": 52,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, _, checks = ACSolver(cryptarithmetic).GAC(arc_heuristic=no_heuristic)\n",
+ "f'GAC needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 58,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 939 ms, sys: 0 ns, total: 939 ms\n",
+ "Wall time: 938 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'GAC with SAT UP arc heuristic needs 573120 consistency-checks'"
+ ]
+ },
+ "execution_count": 58,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, _, checks = ACSolver(cryptarithmetic).GAC(arc_heuristic=sat_up)\n",
+ "f'GAC with SAT UP arc heuristic needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 59,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/latex": [
+ "\\begin{array}{@{}r@{}} 9567 \\\\ + 1085 \\\\ \\hline 10652 \\end{array}"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "assignment = ACSolver(cryptarithmetic).domain_splitting()\n",
+ "\n",
+ "from IPython.display import Latex\n",
+ "display(Latex(r'\\begin{array}{@{}r@{}} ' + '{}{}{}{}'.format(assignment['S'], assignment['E'], assignment['N'], assignment['D']) + r' \\\\ + ' + \n",
+ " '{}{}{}{}'.format(assignment['M'], assignment['O'], assignment['R'], assignment['E']) + r' \\\\ \\hline ' + \n",
+ " '{}{}{}{}{}'.format(assignment['M'], assignment['O'], assignment['N'], assignment['E'], assignment['Y']) + ' \\end{array}'))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "## References\n",
+ "\n",
+ "[[1]](#ref-1) Van Dongen, Marc RC. 2002. _Domain-heuristics for arc-consistency algorithms_.\n",
+ "\n",
+ "[[2]](#ref-2) Van Dongen, MRC and Bowen, JA. 2000. _Improving arc-consistency algorithms with double-support checks_.\n",
+ "\n",
+ "[[3]](#ref-3) Wallace, Richard J and Freuder, Eugene Charles. 1992. _Ordering heuristics for arc consistency algorithms_."
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.7.5rc1"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}
diff --git a/classical_planning_approaches.ipynb b/classical_planning_approaches.ipynb
new file mode 100644
index 000000000..b3373b367
--- /dev/null
+++ b/classical_planning_approaches.ipynb
@@ -0,0 +1,2402 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "# Classical Planning\n",
+ "---\n",
+ "# Classical Planning Approaches\n",
+ "\n",
+ "## Introduction \n",
+ "***Planning*** combines the two major areas of AI: *search* and *logic*. A planner can be seen either as a program that searches for a solution or as one that constructively proves the existence of a solution.\n",
+ "\n",
+ "Currently, the most popular and effective approaches to fully automated planning are:\n",
+ "- searching using a *planning graph*;\n",
+ "- *state-space search* with heuristics;\n",
+ "- translating to a *constraint satisfaction (CSP) problem*;\n",
+ "- translating to a *boolean satisfiability (SAT) problem*."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from planning import *"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Planning as Planning Graph Search\n",
+ "\n",
+ "A *planning graph* is a directed graph organized into levels each of which contains information about the current state of the knowledge base and the possible state-action links to and from that level. \n",
+ "\n",
+ "The first level contains the initial state with nodes representing each fluent that holds in that level. This level has state-action links linking each state to valid actions in that state. Each action is linked to all its preconditions and its effect states. Based on these effects, the next level is constructed and contains similarly structured information about the next state. In this way, the graph is expanded using state-action links till we reach a state where all the required goals hold true simultaneously.\n",
+ "\n",
+ "In every planning problem, we are allowed to carry out the *no-op* action, ie, we can choose no action for a particular state. These are called persistence actions and has effects same as its preconditions. This enables us to carry a state to the next level.\n",
+ "\n",
+ "Mutual exclusivity (*mutex*) between two actions means that these cannot be taken together and occurs in the following cases:\n",
+ "- *inconsistent effects*: one action negates the effect of the other;\n",
+ "- *interference*: one of the effects of an action is the negation of a precondition of the other;\n",
+ "- *competing needs*: one of the preconditions of one action is mutually exclusive with a precondition of the other.\n",
+ "\n",
+ "We can say that we have reached our goal if none of the goal states in the current level are mutually exclusive."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mclass\u001b[0m \u001b[0mGraph\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Contains levels of state and actions\u001b[0m\n",
+ "\u001b[0;34m Used in graph planning algorithm to extract a solution\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mFolKB\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mLevel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mobjects\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0marg\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mclause\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mclauses\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0marg\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__call__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpand_graph\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mexpand_graph\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Expands the graph by a level\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mlast_level\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mlast_level\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mactions\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mobjects\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlast_level\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mperform_actions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mnon_mutex_goals\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindex\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Checks whether the goals are mutually exclusive\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mgoal_perm\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mitertools\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcombinations\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mg\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mgoal_perm\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mg\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource Graph"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mclass\u001b[0m \u001b[0mLevel\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Contains the state of the planning problem\u001b[0m\n",
+ "\u001b[0;34m and exhaustive list of actions which use the\u001b[0m\n",
+ "\u001b[0;34m states as pre-condition.\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Initializes variables to hold state and action details of a level\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mkb\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# current state\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mkb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# current action to state link\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_action_links\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# current state to action link\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state_links\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# current action to next state link\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_action_links\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# next state to current action link\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# mutually exclusive actions\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__call__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mobjects\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbuild\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mactions\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mobjects\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfind_mutex\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mseparate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Separates an iterable of elements into positive and negative parts\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mpositive\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnegative\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mclause\u001b[0m \u001b[0;32min\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'Not'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnegative\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mpositive\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mpositive\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnegative\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mfind_mutex\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Finds mutually exclusive actions\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Inconsistent effects\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mpos_nsl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mneg_nsl\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mseparate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mnegeff\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mneg_nsl\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_negeff\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnegeff\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0mnegeff\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mposeff\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mpos_nsl\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mnew_negeff\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mposeff\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mposeff\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mb\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnegeff\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mb\u001b[0m\u001b[0;34m}\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mb\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Interference will be calculated with the last step\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mpos_csl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mneg_csl\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mseparate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state_links\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Competing needs\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mpos_precond\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mpos_csl\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mneg_precond\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mneg_csl\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_neg_precond\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mneg_precond\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0mneg_precond\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mnew_neg_precond\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mpos_precond\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mpos_precond\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mb\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mneg_precond\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mb\u001b[0m\u001b[0;34m}\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mb\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Inconsistent support\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mstate_mutex\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mpair\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnext_state_0\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpair\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpair\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnext_state_1\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpair\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnext_state_1\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpair\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnext_state_0\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnext_state_1\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mstate_mutex\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0mnext_state_0\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnext_state_1\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mstate_mutex\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mbuild\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mobjects\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Populates the lists and dictionaries containing the state action dependencies\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mclause\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mp_expr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'P'\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mp_expr\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mp_expr\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mp_expr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mp_expr\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnum_args\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mpossible_args\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtuple\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mitertools\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpermutations\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mobjects\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnum_args\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0marg\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mpossible_args\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcheck_precond\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mnum\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msymbol\u001b[0m \u001b[0;32min\u001b[0m \u001b[0menumerate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mislower\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0marg\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0marg\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnum\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0marg\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtuple\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_action\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msubstitute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnew_action\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mclause\u001b[0m \u001b[0;32min\u001b[0m \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprecond\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_clause\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msubstitute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnew_action\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_clause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mnew_clause\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state_links\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnew_clause\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_action\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnew_clause\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mnew_action\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnew_action\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mclause\u001b[0m \u001b[0;32min\u001b[0m \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0meffect\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_clause\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msubstitute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnew_action\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_clause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mnew_clause\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnew_clause\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_action\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnew_clause\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mnew_action\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mperform_actions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Performs the necessary actions and returns a new Level\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_kb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mFolKB\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeys\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mLevel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_kb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource Level"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "A *planning graph* can be used to give better heuristic estimates which can be applied to any of the search techniques. Alternatively, we can search for a solution over the space formed by the planning graph, using an algorithm called `GraphPlan`.\n",
+ "\n",
+ "The `GraphPlan` algorithm repeatedly adds a level to a planning graph. Once all the goals show up as non-mutex in the graph, the algorithm runs backward from the last level to the first searching for a plan that solves the problem. If that fails, it records the (level , goals) pair as a *no-good* (as in constraint learning for CSPs), expands another level and tries again, terminating with failure when there is no reason to go on. "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mclass\u001b[0m \u001b[0mGraphPlan\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Class for formulation GraphPlan algorithm\u001b[0m\n",
+ "\u001b[0;34m Constructs a graph of state and action space\u001b[0m\n",
+ "\u001b[0;34m Returns solution for the planning problem\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mGraph\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mno_goods\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msolution\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mcheck_leveloff\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Checks if the graph has levelled off\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcheck\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_state\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mcheck\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mextract_solution\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindex\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Extracts the solution\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mlevel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnon_mutex_goals\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindex\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mno_goods\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlevel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mlevel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mindex\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Create all combinations of actions that satisfy the goal\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mactions\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mgoal\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlevel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnext_state_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mgoal\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mall_actions\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mitertools\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mproduct\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0mactions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Filter out non-mutex actions\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnon_mutex_actions\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction_tuple\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mall_actions\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0maction_pairs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mitertools\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcombinations\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction_tuple\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnon_mutex_actions\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction_tuple\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mpair\u001b[0m \u001b[0;32min\u001b[0m \u001b[0maction_pairs\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpair\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mlevel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmutex\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnon_mutex_actions\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Recursion\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction_list\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mnon_mutex_actions\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0maction_list\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindex\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msolution\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msolution\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0maction_list\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindex\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_goals\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mact\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction_list\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mact\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mlevel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_action_links\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnew_goals\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnew_goals\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mlevel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcurrent_action_links\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mact\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mabs\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melif\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlevel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnew_goals\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mno_goods\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mextract_solution\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_goals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mindex\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Level-Order multiple solutions\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msolution\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mitem\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msolution\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mitem\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msolution\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msolution\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mitem\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msolution\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mitem\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mnum\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mitem\u001b[0m \u001b[0;32min\u001b[0m \u001b[0menumerate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msolution\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mitem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreverse\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msolution\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnum\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mitem\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0msolution\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mgoal_test\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkb\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkb\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mask\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mq\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mFalse\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mq\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mexecute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Executes the GraphPlan algorithm for the given problem\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpand_graph\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoal_test\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkb\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnon_mutex_goals\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msolution\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mextract_solution\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0msolution\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0msolution\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgraph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlevels\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>=\u001b[0m \u001b[0;36m2\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcheck_leveloff\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource GraphPlan"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Planning as State-Space Search"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "The description of a planning problem defines a search problem: we can search from the initial state through the space of states, looking for a goal. One of the nice advantages of the declarative representation of action schemas is that we can also search backward from the goal, looking for the initial state. \n",
+ "\n",
+ "However, neither forward nor backward search is efficient without a good heuristic function because the real-world planning problems often have large state spaces. A heuristic function $h(s)$ estimates the distance from a state $s$ to the goal and, if it is admissible, ie if does not overestimate, then we can use $A^∗$ search to find optimal solutions.\n",
+ "\n",
+ "Planning uses a factored representation for states and action schemas which makes it possible to define good domain-independent heuristics to prune the search space.\n",
+ "\n",
+ "An admissible heuristic can be derived by defining a relaxed problem that is easier to solve. The length of the solution of this easier problem then becomes the heuristic for the original problem. Assume that all goals and preconditions contain only positive literals, ie that the problem is defined according to the *Stanford Research Institute Problem Solver* (STRIPS) notation: we want to create a relaxed version of the original problem that will be easier to solve by ignoring delete lists from all actions, ie removing all negative literals from effects. As shown in [[1]](#cite-hoffmann2001ff) the planning graph of a relaxed problem does not contain any mutex relations at all (which is the crucial thing when building a planning graph) and for this reason GraphPlan will never backtrack looking for a solution: for this reason the **ignore delete lists** heuristic makes it possible to find the optimal solution for relaxed problem in polynomial time through `GraphPlan` algorithm."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from search import *"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Forward State-Space Search"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Forward search through the space of states, starting in the initial state and using the problem’s actions to search forward for a member of the set of goal states."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mclass\u001b[0m \u001b[0mForwardPlan\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msearch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mProblem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m [Section 10.2.1]\u001b[0m\n",
+ "\u001b[0;34m Forward state-space search\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msuper\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpanded_actions\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpand_actions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0maction\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpanded_actions\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpre\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mpre\u001b[0m \u001b[0;32min\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprecond\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mresult\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mgoal_test\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgoal\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mgoal\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mh\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Computes ignore delete lists heuristic by creating a relaxed version of the original problem (we can do that\u001b[0m\n",
+ "\u001b[0;34m by removing the delete lists from all actions, i.e. removing all negative literals from effects) that will be\u001b[0m\n",
+ "\u001b[0;34m easier to solve through GraphPlan and where the length of the solution will serve as a good heuristic.\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrelaxed_planning_problem\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mPlanningProblem\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoal\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrelaxed\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32min\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mactions\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlinearize\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mGraphPlan\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrelaxed_planning_problem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexecute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mexcept\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mfloat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'inf'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource ForwardPlan"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Backward Relevant-States Search"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Backward search through sets of relevant states, starting at the set of states representing the goal and using the inverse of the actions to search backward for the initial state."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mclass\u001b[0m \u001b[0mBackwardPlan\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msearch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mProblem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m [Section 10.2.2]\u001b[0m\n",
+ "\u001b[0;34m Backward relevant-states search\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msuper\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpanded_actions\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpand_actions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msubgoal\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Returns True if the action is relevant to the subgoal, i.e.:\u001b[0m\n",
+ "\u001b[0;34m - the action achieves an element of the effects\u001b[0m\n",
+ "\u001b[0;34m - the action doesn't delete something that needs to be achieved\u001b[0m\n",
+ "\u001b[0;34m - the preconditions are consistent with other subgoals that need to be achieved\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mnegate_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreplace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Not'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m''\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'Not'\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m'Not'\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msubgoal\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msubgoal\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0maction\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpanded_actions\u001b[0m \u001b[0;32mif\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0many\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprop\u001b[0m \u001b[0;32min\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0meffect\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mprop\u001b[0m \u001b[0;32min\u001b[0m \u001b[0msubgoal\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mand\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0many\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnegate_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprop\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0msubgoal\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mprop\u001b[0m \u001b[0;32min\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0meffect\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mand\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0many\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnegate_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprop\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0msubgoal\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mnegate_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprop\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0meffect\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mprop\u001b[0m \u001b[0;32min\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprecond\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mresult\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msubgoal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# g' = (g - effects(a)) + preconds(a)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msubgoal\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdifference\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0meffect\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0munion\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprecond\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mgoal_test\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msubgoal\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mall\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgoal\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoal\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mgoal\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msubgoal\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mh\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msubgoal\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Computes ignore delete lists heuristic by creating a relaxed version of the original problem (we can do that\u001b[0m\n",
+ "\u001b[0;34m by removing the delete lists from all actions, i.e. removing all negative literals from effects) that will be\u001b[0m\n",
+ "\u001b[0;34m easier to solve through GraphPlan and where the length of the solution will serve as a good heuristic.\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrelaxed_planning_problem\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mPlanningProblem\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoal\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msubgoal\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrelaxed\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32min\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mactions\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlinearize\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mGraphPlan\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrelaxed_planning_problem\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexecute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mexcept\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mfloat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'inf'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource BackwardPlan"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Planning as Constraint Satisfaction Problem"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In forward planning, the search is constrained by the initial state and only uses the goal as a stopping criterion and as a source for heuristics. In regression planning, the search is constrained by the goal and only uses the start state as a stopping criterion and as a source for heuristics. By converting the problem to a constraint satisfaction problem (CSP), the initial state can be used to prune what is not reachable and the goal to prune what is not useful. The CSP will be defined for a finite number of steps; the number of steps can be adjusted to find the shortest plan. One of the CSP methods can then be used to solve the CSP and thus find a plan.\n",
+ "\n",
+ "To construct a CSP from a planning problem, first choose a fixed planning *horizon*, which is the number of time steps over which to plan. Suppose the horizon is \n",
+ "$k$. The CSP has the following variables:\n",
+ "\n",
+ "- a *state variable* for each feature and each time from 0 to $k$. If there are $n$ features for a horizon of $k$, there are $n \\cdot (k+1)$ state variables. The domain of the state variable is the domain of the corresponding feature;\n",
+ "- an *action variable*, $Action_t$, for each $t$ in the range 0 to $k-1$. The domain of $Action_t$, represents the action that takes the agent from the state at time $t$ to the state at time $t+1$.\n",
+ "\n",
+ "There are several types of constraints:\n",
+ "\n",
+ "- a *precondition constraint* between a state variable at time $t$ and the variable $Actiont_t$ constrains what actions are legal at time $t$;\n",
+ "- an *effect constraint* between $Action_t$ and a state variable at time $t+1$ constrains the values of a state variable that is a direct effect of the action;\n",
+ "- a *frame constraint* among a state variable at time $t$, the variable $Action_t$, and the corresponding state variable at time $t+1$ specifies when the variable that does not change as a result of an action has the same value before and after the action;\n",
+ "- an *initial-state constraint* constrains a variable on the initial state (at time 0). The initial state is represented as a set of domain constraints on the state variables at time 0;\n",
+ "- a *goal constraint* constrains the final state to be a state that satisfies the achievement goal. These are domain constraints on the variables that appear in the goal;\n",
+ "- a *state constraint* is a constraint among variables at the same time step. These can include physical constraints on the state or can ensure that states that violate maintenance goals are forbidden. This is extra knowledge beyond the power of the feature-based or PDDL representations of the action.\n",
+ "\n",
+ "The PDDL representation gives precondition, effect and frame constraints for each time \n",
+ "$t$ as follows:\n",
+ "\n",
+ "- for each $Var = v$ in the precondition of action $A$, there is a precondition constraint:\n",
+ "$$ Var_t = v \\leftarrow Action_t = A $$\n",
+ "that specifies that if the action is to be $A$, $Var_t$ must have value $v$ immediately before. This constraint is violated when $Action_t = A$ and $Var_t \\neq v$, and thus is equivalent to $\\lnot{(Var_t \\neq v \\land Action_t = A)}$;\n",
+ "- or each $Var = v$ in the effect of action $A$, there is a effect constraint:\n",
+ "$$ Var_{t+1} = v \\leftarrow Action_t = A $$\n",
+ "which is violated when $Action_t = A$ and $Var_{t+1} \\neq v$, and thus is equivalent to $\\lnot{(Var_{t+1} \\neq v \\land Action_t = A)}$;\n",
+ "- for each $Var$, there is a frame constraint, where $As$ is the set of actions that include $Var$ in the effect of the action:\n",
+ "$$ Var_{t+1} = Var_t \\leftarrow Action_t \\notin As $$\n",
+ "which specifies that the feature $Var$ has the same value before and after any action that does not affect $Var$.\n",
+ "\n",
+ "The CSP representation assumes a fixed planning horizon (ie a fixed number of steps). To find a plan over any number of steps, the algorithm can be run for a horizon of $k = 0, 1, 2, \\dots$ until a solution is found."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from csp import *"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mCSPlan\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msolution_length\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mCSP_solver\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mac_search_solver\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0msat_up\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m [Section 10.4.3]\u001b[0m\n",
+ "\u001b[0;34m Planning as Constraint Satisfaction Problem\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Returns a string for the var-stage pair that can be used as a variable\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m\"_\"\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mif_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mv1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mv2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"If the second argument is v2, the first argument must be v1\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mif_fun\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mx1\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mv1\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mx2\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mv2\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mif_fun\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"if the second argument is \"\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mv2\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m\" then the first argument is \"\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mv1\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m\" \"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mif_fun\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0meq_if_not_in_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mactset\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"First and third arguments are equal if action is not in actset\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0meq_if_not_in\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0ma\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mx1\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mx2\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0ma\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mactset\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meq_if_not_in\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__name__\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"first and third arguments are equal if action is not in \"\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mactset\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;34m\" \"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0meq_if_not_in\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mexpanded_actions\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpand_actions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mfluent_values\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpand_fluents\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mhorizon\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msolution_length\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mact_vars\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'action'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mstage\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhorizon\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomains\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mav\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mlist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexpanded_actions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mav\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mact_vars\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomains\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mupdate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m{\u001b[0m\u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m}\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvar\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mfluent_values\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mstage\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhorizon\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# initial state constraints\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconstraints\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mConstraint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mval\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfluent\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreplace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Not'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m''\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfluent\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m'Not'\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mfluent\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconstraints\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mConstraint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvar\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfluent\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreplace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Not'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m''\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mfluent\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mfluent_values\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfluent\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# goal state constraints\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconstraints\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mConstraint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhorizon\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mval\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfluent\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreplace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Not'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m''\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfluent\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m'Not'\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mfluent\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# precondition constraints\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconstraints\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mConstraint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'action'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mif_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mact\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# st(var, stage) == val if st('action', stage) == act\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mact\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrps\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mexpanded_actions\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mval\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfluent\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreplace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Not'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m''\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfluent\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m'Not'\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mfluent\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstrps\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprecond\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mstage\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhorizon\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# effect constraints\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconstraints\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mConstraint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'action'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mif_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mval\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mact\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# st(var, stage + 1) == val if st('action', stage) == act\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mact\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrps\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mexpanded_actions\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mval\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfluent\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreplace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Not'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m''\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfluent\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m'Not'\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mfluent\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstrps\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0meffect\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mstage\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhorizon\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# frame constraints\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconstraints\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mConstraint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'action'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstage\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meq_if_not_in_\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mact\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mact\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mexpanded_actions\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mvar\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mact\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0meffect\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Not'\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mvar\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0mvar\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mact\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0meffect\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mvar\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mfluent_values\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mstage\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhorizon\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcsp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mNaryCSP\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdomains\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mconstraints\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msol\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCSP_solver\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcsp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0marc_heuristic\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0marc_heuristic\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0msol\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0msol\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ma\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ma\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mact_vars\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource CSPlan"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Planning as Boolean Satisfiability Problem\n",
+ "\n",
+ "As shown in [[2]](cite-kautz1992planning) the translation of a *Planning Domain Definition Language* (PDDL) description into a *Conjunctive Normal Form* (CNF) formula is a series of straightforward steps:\n",
+ "- *propositionalize the actions*: replace each action schema with a set of ground actions formed by substituting constants for each of the variables. These ground actions are not part of the translation, but will be used in subsequent steps;\n",
+ "- *define the initial state*: assert $F^0$ for every fluent $F$ in the problem’s initial state, and $\\lnot{F}$ for every fluent not mentioned in the initial state;\n",
+ "- *propositionalize the goal*: for every variable in the goal, replace the literals that contain the variable with a disjunction over constants;\n",
+ "- *add successor-state axioms*: for each fluent $F$, add an axiom of the form\n",
+ "\n",
+ "$$ F^{t+1} \\iff ActionCausesF^t \\lor (F^t \\land \\lnot{ActionCausesNotF^t}) $$\n",
+ "\n",
+ "where $ActionCausesF$ is a disjunction of all the ground actions that have $F$ in their add list, and $ActionCausesNotF$ is a disjunction of all the ground actions that have $F$ in their delete list;\n",
+ "- *add precondition axioms*: for each ground action $A$, add the axiom $A^t \\implies PRE(A)^t$, that is, if an action is taken at time $t$, then the preconditions must have been true;\n",
+ "- *add action exclusion axioms*: say that every action is distinct from every other action.\n",
+ "\n",
+ "A propositional planning procedure implements the basic idea just given but, because the agent does not know how many steps it will take to reach the goal, the algorithm tries each possible number of steps $t$, up to some maximum conceivable plan length $T_{max}$ . In this way, it is guaranteed to find the shortest plan if one exists. Because of the way the propositional planning procedure searches for a solution, this approach cannot be used in a partially observable environment, ie WalkSAT, but would just set the unobservable variables to the values it needs to create a solution."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from logic import *"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mSATPlan\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msolution_length\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSAT_solver\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcdcl_satisfiable\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m [Section 10.4.1]\u001b[0m\n",
+ "\u001b[0;34m Planning as Boolean satisfiability\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mexpand_transitions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mstate\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msorted\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mact\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mact\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcheck_precond\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mact\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mupdate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msorted\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfilter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0;34m'Not'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_strips\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msorted\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mstate\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mstate\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mexpand_transitions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mexpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mtransition\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdefaultdict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdict\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mexpand_transitions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexpand_actions\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mSAT_plan\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msorted\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msorted\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mplanning_problem\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgoals\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msolution_length\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSAT_solver\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mSAT_solver\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource SATPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mSAT_plan\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minit\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgoal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt_max\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mSAT_solver\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcdcl_satisfiable\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Converts a planning problem to Satisfaction problem by translating it to a cnf sentence.\u001b[0m\n",
+ "\u001b[0;34m [Figure 7.22]\u001b[0m\n",
+ "\u001b[0;34m >>> transition = {'A': {'Left': 'A', 'Right': 'B'}, 'B': {'Left': 'A', 'Right': 'C'}, 'C': {'Left': 'B', 'Right': 'C'}}\u001b[0m\n",
+ "\u001b[0;34m >>> SAT_plan('A', transition, 'C', 1) is None\u001b[0m\n",
+ "\u001b[0;34m True\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Functions used by SAT_plan\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mtranslate_to_SAT\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minit\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgoal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mstates\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mstate\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mstate\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Symbol claiming state s at time t\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mstate_counter\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mitertools\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcount\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstates\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mt\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtime\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mstate_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"S{}\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate_counter\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Add initial state axiom\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0minit\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Add goal state axiom\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mfirst\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mclause\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstate_sym\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0missuperset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgoal\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \\\n",
+ " \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mgoal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mExpr\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstate_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mgoal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# All possible transitions\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mtransition_counter\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mitertools\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcount\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstates\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0maction\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0ms_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0maction\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mt\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Action 'action' taken from state 's' at time 't' to reach 's_'\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0maction_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mExpr\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"T{}\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtransition_counter\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Change the state from s to s_\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m|\u001b[0m \u001b[0;34m'==>'\u001b[0m \u001b[0;34m|\u001b[0m \u001b[0mstate_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maction_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m|\u001b[0m \u001b[0;34m'==>'\u001b[0m \u001b[0;34m|\u001b[0m \u001b[0mstate_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Allow only one state at any time\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mt\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtime\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# must be a state at any time\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'|'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mstate_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstates\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstates\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms_\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mstates\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mstates\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# for each pair of states s, s_ only one is possible at time t\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0mstate_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m|\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0mstate_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ms_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Restrict to one transition per timestep\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mt\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# list of possible transitions at time t\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mtransitions_t\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mtr\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mtr\u001b[0m \u001b[0;32min\u001b[0m \u001b[0maction_sym\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mtr\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# make sure at least one of the transitions happens\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'|'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0maction_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mtr\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mtr\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtransitions_t\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mtr\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtransitions_t\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mtr_\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtransitions_t\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mtransitions_t\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtr\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# there cannot be two transitions tr and tr_ at time t\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0maction_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mtr\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m|\u001b[0m \u001b[0;34m~\u001b[0m\u001b[0maction_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mtr_\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Combine the clauses to form the cnf\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'&'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mextract_solution\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mtrue_transitions\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mt\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mt\u001b[0m \u001b[0;32min\u001b[0m \u001b[0maction_sym\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0maction_sym\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mt\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Sort transitions based on time, which is the 3rd element of the tuple\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mtrue_transitions\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msort\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0maction\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0maction\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtime\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtrue_transitions\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# Body of SAT_plan algorithm\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mt\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mt_max\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# dictionaries to help extract the solution from model\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mstate_sym\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0maction_sym\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mcnf\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtranslate_to_SAT\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minit\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtransition\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mgoal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mSAT_solver\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcnf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mmodel\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mextract_solution\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource SAT_plan"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "## Experimental Results"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Blocks World"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 13,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mthree_block_tower\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m [Figure 10.3] THREE-BLOCK-TOWER\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m A blocks-world problem of stacking three blocks in a certain configuration,\u001b[0m\n",
+ "\u001b[0;34m also known as the Sussman Anomaly.\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m Example:\u001b[0m\n",
+ "\u001b[0;34m >>> from planning import *\u001b[0m\n",
+ "\u001b[0;34m >>> tbt = three_block_tower()\u001b[0m\n",
+ "\u001b[0;34m >>> tbt.goal_test()\u001b[0m\n",
+ "\u001b[0;34m False\u001b[0m\n",
+ "\u001b[0;34m >>> tbt.act(expr('MoveToTable(C, A)'))\u001b[0m\n",
+ "\u001b[0;34m >>> tbt.act(expr('Move(B, Table, C)'))\u001b[0m\n",
+ "\u001b[0;34m >>> tbt.goal_test()\u001b[0m\n",
+ "\u001b[0;34m False\u001b[0m\n",
+ "\u001b[0;34m >>> tbt.act(expr('Move(A, Table, B)'))\u001b[0m\n",
+ "\u001b[0;34m >>> tbt.goal_test()\u001b[0m\n",
+ "\u001b[0;34m True\u001b[0m\n",
+ "\u001b[0;34m >>>\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mPlanningProblem\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'On(A, Table) & On(B, Table) & On(C, A) & Clear(B) & Clear(C)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'On(A, B) & On(B, C)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Move(b, x, y)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'On(b, x) & Clear(b) & Clear(y)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'On(b, y) & Clear(x) & ~On(b, x) & ~Clear(y)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Block(b) & Block(y)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'MoveToTable(b, x)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'On(b, x) & Clear(b)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'On(b, Table) & Clear(x) & ~On(b, x)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Block(b) & Block(x)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Block(A) & Block(B) & Block(C)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource three_block_tower"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### GraphPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 4.46 ms, sys: 124 µs, total: 4.59 ms\n",
+ "Wall time: 4.48 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[MoveToTable(C, A), Move(B, Table, C), Move(A, Table, B)]"
+ ]
+ },
+ "execution_count": 14,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time blocks_world_solution = GraphPlan(three_block_tower()).execute()\n",
+ "linearize(blocks_world_solution)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### ForwardPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 15,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "14 paths have been expanded and 28 paths remain in the frontier\n",
+ "CPU times: user 91 ms, sys: 0 ns, total: 91 ms\n",
+ "Wall time: 89.8 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[MoveToTable(C, A), Move(B, Table, C), Move(A, Table, B)]"
+ ]
+ },
+ "execution_count": 15,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time blocks_world_solution = uniform_cost_search(ForwardPlan(three_block_tower()), display=True).solution()\n",
+ "blocks_world_solution = list(map(lambda action: Expr(action.name, *action.args), blocks_world_solution))\n",
+ "blocks_world_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### ForwardPlan with Ignore Delete Lists Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 16,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "3 paths have been expanded and 9 paths remain in the frontier\n",
+ "CPU times: user 81.3 ms, sys: 3.11 ms, total: 84.5 ms\n",
+ "Wall time: 83 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[MoveToTable(C, A), Move(B, Table, C), Move(A, Table, B)]"
+ ]
+ },
+ "execution_count": 16,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time blocks_world_solution = astar_search(ForwardPlan(three_block_tower()), display=True).solution()\n",
+ "blocks_world_solution = list(map(lambda action: Expr(action.name, *action.args), blocks_world_solution))\n",
+ "blocks_world_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### BackwardPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 17,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "116 paths have been expanded and 289 paths remain in the frontier\n",
+ "CPU times: user 266 ms, sys: 718 µs, total: 267 ms\n",
+ "Wall time: 265 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[MoveToTable(C, A), Move(B, Table, C), Move(A, Table, B)]"
+ ]
+ },
+ "execution_count": 17,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time blocks_world_solution = uniform_cost_search(BackwardPlan(three_block_tower()), display=True).solution()\n",
+ "blocks_world_solution = list(map(lambda action: Expr(action.name, *action.args), blocks_world_solution))\n",
+ "blocks_world_solution[::-1]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### BackwardPlan with Ignore Delete Lists Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 18,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "4 paths have been expanded and 20 paths remain in the frontier\n",
+ "CPU times: user 477 ms, sys: 450 µs, total: 477 ms\n",
+ "Wall time: 476 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[MoveToTable(C, A), Move(B, Table, C), Move(A, Table, B)]"
+ ]
+ },
+ "execution_count": 18,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time blocks_world_solution = astar_search(BackwardPlan(three_block_tower()), display=True).solution()\n",
+ "blocks_world_solution = list(map(lambda action: Expr(action.name, *action.args), blocks_world_solution))\n",
+ "blocks_world_solution[::-1]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 172 ms, sys: 4.52 ms, total: 176 ms\n",
+ "Wall time: 175 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[MoveToTable(C, A), Move(B, Table, C), Move(A, Table, B)]"
+ ]
+ },
+ "execution_count": 19,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time blocks_world_solution = CSPlan(three_block_tower(), 3, arc_heuristic=no_heuristic)\n",
+ "blocks_world_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSPlan with SAT UP Arc Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 267 ms, sys: 0 ns, total: 267 ms\n",
+ "Wall time: 266 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[MoveToTable(C, A), Move(B, Table, C), Move(A, Table, B)]"
+ ]
+ },
+ "execution_count": 20,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time blocks_world_solution = CSPlan(three_block_tower(), 3, arc_heuristic=sat_up)\n",
+ "blocks_world_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### SATPlan with DPLL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 27,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 34.9 s, sys: 15.9 ms, total: 34.9 s\n",
+ "Wall time: 34.9 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[MoveToTable(C, A), Move(B, Table, C), Move(A, Table, B)]"
+ ]
+ },
+ "execution_count": 27,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time blocks_world_solution = SATPlan(three_block_tower(), 4, SAT_solver=dpll_satisfiable)\n",
+ "blocks_world_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### SATPlan with CDCL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 1.15 s, sys: 4.01 ms, total: 1.15 s\n",
+ "Wall time: 1.15 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[MoveToTable(C, A), Move(B, Table, C), Move(A, Table, B)]"
+ ]
+ },
+ "execution_count": 28,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time blocks_world_solution = SATPlan(three_block_tower(), 4, SAT_solver=cdcl_satisfiable)\n",
+ "blocks_world_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Spare Tire"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 21,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mspare_tire\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m [Figure 10.2] SPARE-TIRE-PROBLEM\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m A problem involving changing the flat tire of a car\u001b[0m\n",
+ "\u001b[0;34m with a spare tire from the trunk.\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m Example:\u001b[0m\n",
+ "\u001b[0;34m >>> from planning import *\u001b[0m\n",
+ "\u001b[0;34m >>> st = spare_tire()\u001b[0m\n",
+ "\u001b[0;34m >>> st.goal_test()\u001b[0m\n",
+ "\u001b[0;34m False\u001b[0m\n",
+ "\u001b[0;34m >>> st.act(expr('Remove(Spare, Trunk)'))\u001b[0m\n",
+ "\u001b[0;34m >>> st.act(expr('Remove(Flat, Axle)'))\u001b[0m\n",
+ "\u001b[0;34m >>> st.goal_test()\u001b[0m\n",
+ "\u001b[0;34m False\u001b[0m\n",
+ "\u001b[0;34m >>> st.act(expr('PutOn(Spare, Axle)'))\u001b[0m\n",
+ "\u001b[0;34m >>> st.goal_test()\u001b[0m\n",
+ "\u001b[0;34m True\u001b[0m\n",
+ "\u001b[0;34m >>>\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mPlanningProblem\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(Flat, Axle) & At(Spare, Trunk)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(Spare, Axle) & At(Flat, Ground)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Remove(obj, loc)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(obj, loc)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(obj, Ground) & ~At(obj, loc)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Tire(obj)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'PutOn(t, Axle)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(t, Ground) & ~At(Flat, Axle)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(t, Axle) & ~At(t, Ground)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Tire(t)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'LeaveOvernight'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m''\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'~At(Spare, Ground) & ~At(Spare, Axle) & ~At(Spare, Trunk) & \\\u001b[0m\n",
+ "\u001b[0;34m ~At(Flat, Ground) & ~At(Flat, Axle) & ~At(Flat, Trunk)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Tire(Flat) & Tire(Spare)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource spare_tire"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### GraphPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 29,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 4.24 ms, sys: 1 µs, total: 4.24 ms\n",
+ "Wall time: 4.16 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Remove(Flat, Axle), Remove(Spare, Trunk), PutOn(Spare, Axle)]"
+ ]
+ },
+ "execution_count": 29,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time spare_tire_solution = GraphPlan(spare_tire()).execute()\n",
+ "linearize(spare_tire_solution)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### ForwardPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 30,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "11 paths have been expanded and 9 paths remain in the frontier\n",
+ "CPU times: user 10.3 ms, sys: 0 ns, total: 10.3 ms\n",
+ "Wall time: 9.89 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Remove(Flat, Axle), Remove(Spare, Trunk), PutOn(Spare, Axle)]"
+ ]
+ },
+ "execution_count": 30,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time spare_tire_solution = uniform_cost_search(ForwardPlan(spare_tire()), display=True).solution()\n",
+ "spare_tire_solution = list(map(lambda action: Expr(action.name, *action.args), spare_tire_solution))\n",
+ "spare_tire_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### ForwardPlan with Ignore Delete Lists Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 31,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "5 paths have been expanded and 8 paths remain in the frontier\n",
+ "CPU times: user 20.4 ms, sys: 1 µs, total: 20.4 ms\n",
+ "Wall time: 19.4 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Remove(Flat, Axle), Remove(Spare, Trunk), PutOn(Spare, Axle)]"
+ ]
+ },
+ "execution_count": 31,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time spare_tire_solution = astar_search(ForwardPlan(spare_tire()), display=True).solution()\n",
+ "spare_tire_solution = list(map(lambda action: Expr(action.name, *action.args), spare_tire_solution))\n",
+ "spare_tire_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### BackwardPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 32,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "29 paths have been expanded and 22 paths remain in the frontier\n",
+ "CPU times: user 22.2 ms, sys: 7 µs, total: 22.2 ms\n",
+ "Wall time: 21.3 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Remove(Flat, Axle), Remove(Spare, Trunk), PutOn(Spare, Axle)]"
+ ]
+ },
+ "execution_count": 32,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time spare_tire_solution = uniform_cost_search(BackwardPlan(spare_tire()), display=True).solution()\n",
+ "spare_tire_solution = list(map(lambda action: Expr(action.name, *action.args), spare_tire_solution))\n",
+ "spare_tire_solution[::-1]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### BackwardPlan with Ignore Delete Lists Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 33,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "3 paths have been expanded and 11 paths remain in the frontier\n",
+ "CPU times: user 13 ms, sys: 0 ns, total: 13 ms\n",
+ "Wall time: 12.5 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Remove(Spare, Trunk), Remove(Flat, Axle), PutOn(Spare, Axle)]"
+ ]
+ },
+ "execution_count": 33,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time spare_tire_solution = astar_search(BackwardPlan(spare_tire()), display=True).solution()\n",
+ "spare_tire_solution = list(map(lambda action: Expr(action.name, *action.args), spare_tire_solution))\n",
+ "spare_tire_solution[::-1]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 34,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 94.7 ms, sys: 0 ns, total: 94.7 ms\n",
+ "Wall time: 93.2 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Remove(Spare, Trunk), Remove(Flat, Axle), PutOn(Spare, Axle)]"
+ ]
+ },
+ "execution_count": 34,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time spare_tire_solution = CSPlan(spare_tire(), 3, arc_heuristic=no_heuristic)\n",
+ "spare_tire_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSPlan with SAT UP Arc Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 35,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 119 ms, sys: 0 ns, total: 119 ms\n",
+ "Wall time: 118 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Remove(Spare, Trunk), Remove(Flat, Axle), PutOn(Spare, Axle)]"
+ ]
+ },
+ "execution_count": 35,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time spare_tire_solution = CSPlan(spare_tire(), 3, arc_heuristic=sat_up)\n",
+ "spare_tire_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### SATPlan with DPLL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 36,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 9.01 s, sys: 3.98 ms, total: 9.01 s\n",
+ "Wall time: 9.01 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Remove(Flat, Axle), Remove(Spare, Trunk), PutOn(Spare, Axle)]"
+ ]
+ },
+ "execution_count": 36,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time spare_tire_solution = SATPlan(spare_tire(), 4, SAT_solver=dpll_satisfiable)\n",
+ "spare_tire_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### SATPlan with CDCL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 37,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 630 ms, sys: 6 µs, total: 630 ms\n",
+ "Wall time: 628 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Remove(Spare, Trunk), Remove(Flat, Axle), PutOn(Spare, Axle)]"
+ ]
+ },
+ "execution_count": 37,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time spare_tire_solution = SATPlan(spare_tire(), 4, SAT_solver=cdcl_satisfiable)\n",
+ "spare_tire_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Shopping Problem"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mshopping_problem\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m SHOPPING-PROBLEM\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m A problem of acquiring some items given their availability at certain stores.\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m Example:\u001b[0m\n",
+ "\u001b[0;34m >>> from planning import *\u001b[0m\n",
+ "\u001b[0;34m >>> sp = shopping_problem()\u001b[0m\n",
+ "\u001b[0;34m >>> sp.goal_test()\u001b[0m\n",
+ "\u001b[0;34m False\u001b[0m\n",
+ "\u001b[0;34m >>> sp.act(expr('Go(Home, HW)'))\u001b[0m\n",
+ "\u001b[0;34m >>> sp.act(expr('Buy(Drill, HW)'))\u001b[0m\n",
+ "\u001b[0;34m >>> sp.act(expr('Go(HW, SM)'))\u001b[0m\n",
+ "\u001b[0;34m >>> sp.act(expr('Buy(Banana, SM)'))\u001b[0m\n",
+ "\u001b[0;34m >>> sp.goal_test()\u001b[0m\n",
+ "\u001b[0;34m False\u001b[0m\n",
+ "\u001b[0;34m >>> sp.act(expr('Buy(Milk, SM)'))\u001b[0m\n",
+ "\u001b[0;34m >>> sp.goal_test()\u001b[0m\n",
+ "\u001b[0;34m True\u001b[0m\n",
+ "\u001b[0;34m >>>\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mPlanningProblem\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(Home) & Sells(SM, Milk) & Sells(SM, Banana) & Sells(HW, Drill)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Have(Milk) & Have(Banana) & Have(Drill)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Buy(x, store)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(store) & Sells(store, x)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Have(x)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Store(store) & Item(x)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Go(x, y)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(x)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(y) & ~At(x)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Place(x) & Place(y)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Place(Home) & Place(SM) & Place(HW) & Store(SM) & Store(HW) & '\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m'Item(Milk) & Item(Banana) & Item(Drill)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource shopping_problem"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### GraphPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 45,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 5.08 ms, sys: 3 µs, total: 5.08 ms\n",
+ "Wall time: 5.03 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Go(Home, HW), Go(Home, SM), Buy(Milk, SM), Buy(Drill, HW), Buy(Banana, SM)]"
+ ]
+ },
+ "execution_count": 45,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time shopping_problem_solution = GraphPlan(shopping_problem()).execute()\n",
+ "linearize(shopping_problem_solution)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### ForwardPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 46,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "167 paths have been expanded and 257 paths remain in the frontier\n",
+ "CPU times: user 187 ms, sys: 4.01 ms, total: 191 ms\n",
+ "Wall time: 190 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Go(Home, SM), Buy(Banana, SM), Buy(Milk, SM), Go(SM, HW), Buy(Drill, HW)]"
+ ]
+ },
+ "execution_count": 46,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time shopping_problem_solution = uniform_cost_search(ForwardPlan(shopping_problem()), display=True).solution()\n",
+ "shopping_problem_solution = list(map(lambda action: Expr(action.name, *action.args), shopping_problem_solution))\n",
+ "shopping_problem_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### ForwardPlan with Ignore Delete Lists Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 47,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "9 paths have been expanded and 22 paths remain in the frontier\n",
+ "CPU times: user 101 ms, sys: 3 µs, total: 101 ms\n",
+ "Wall time: 100 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Go(Home, SM), Buy(Banana, SM), Buy(Milk, SM), Go(SM, HW), Buy(Drill, HW)]"
+ ]
+ },
+ "execution_count": 47,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time shopping_problem_solution = astar_search(ForwardPlan(shopping_problem()), display=True).solution()\n",
+ "shopping_problem_solution = list(map(lambda action: Expr(action.name, *action.args), shopping_problem_solution))\n",
+ "shopping_problem_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### BackwardPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 48,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "176 paths have been expanded and 7 paths remain in the frontier\n",
+ "CPU times: user 109 ms, sys: 2 µs, total: 109 ms\n",
+ "Wall time: 107 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Go(Home, HW), Buy(Drill, HW), Go(HW, SM), Buy(Milk, SM), Buy(Banana, SM)]"
+ ]
+ },
+ "execution_count": 48,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time shopping_problem_solution = uniform_cost_search(BackwardPlan(shopping_problem()), display=True).solution()\n",
+ "shopping_problem_solution = list(map(lambda action: Expr(action.name, *action.args), shopping_problem_solution))\n",
+ "shopping_problem_solution[::-1]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### BackwardPlan with Ignore Delete Lists Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 49,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "18 paths have been expanded and 28 paths remain in the frontier\n",
+ "CPU times: user 235 ms, sys: 9 µs, total: 235 ms\n",
+ "Wall time: 234 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Go(Home, SM), Buy(Banana, SM), Buy(Milk, SM), Go(SM, HW), Buy(Drill, HW)]"
+ ]
+ },
+ "execution_count": 49,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time shopping_problem_solution = astar_search(BackwardPlan(shopping_problem()), display=True).solution()\n",
+ "shopping_problem_solution = list(map(lambda action: Expr(action.name, *action.args), shopping_problem_solution))\n",
+ "shopping_problem_solution[::-1]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 50,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 194 ms, sys: 6 µs, total: 194 ms\n",
+ "Wall time: 192 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Go(Home, HW), Buy(Drill, HW), Go(HW, SM), Buy(Banana, SM), Buy(Milk, SM)]"
+ ]
+ },
+ "execution_count": 50,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time shopping_problem_solution = CSPlan(shopping_problem(), 5, arc_heuristic=no_heuristic)\n",
+ "shopping_problem_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSPlan with SAT UP Arc Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 51,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 235 ms, sys: 7 µs, total: 235 ms\n",
+ "Wall time: 233 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Go(Home, HW), Buy(Drill, HW), Go(HW, SM), Buy(Banana, SM), Buy(Milk, SM)]"
+ ]
+ },
+ "execution_count": 51,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time shopping_problem_solution = CSPlan(shopping_problem(), 5, arc_heuristic=sat_up)\n",
+ "shopping_problem_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### SATPlan with CDCL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 1min 29s, sys: 36 ms, total: 1min 29s\n",
+ "Wall time: 1min 29s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Go(Home, HW), Buy(Drill, HW), Go(HW, SM), Buy(Banana, SM), Buy(Milk, SM)]"
+ ]
+ },
+ "execution_count": 11,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time shopping_problem_solution = SATPlan(shopping_problem(), 5, SAT_solver=cdcl_satisfiable)\n",
+ "shopping_problem_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Air Cargo"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mair_cargo\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m [Figure 10.1] AIR-CARGO-PROBLEM\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m An air-cargo shipment problem for delivering cargo to different locations,\u001b[0m\n",
+ "\u001b[0;34m given the starting location and airplanes.\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m Example:\u001b[0m\n",
+ "\u001b[0;34m >>> from planning import *\u001b[0m\n",
+ "\u001b[0;34m >>> ac = air_cargo()\u001b[0m\n",
+ "\u001b[0;34m >>> ac.goal_test()\u001b[0m\n",
+ "\u001b[0;34m False\u001b[0m\n",
+ "\u001b[0;34m >>> ac.act(expr('Load(C2, P2, JFK)'))\u001b[0m\n",
+ "\u001b[0;34m >>> ac.act(expr('Load(C1, P1, SFO)'))\u001b[0m\n",
+ "\u001b[0;34m >>> ac.act(expr('Fly(P1, SFO, JFK)'))\u001b[0m\n",
+ "\u001b[0;34m >>> ac.act(expr('Fly(P2, JFK, SFO)'))\u001b[0m\n",
+ "\u001b[0;34m >>> ac.act(expr('Unload(C2, P2, SFO)'))\u001b[0m\n",
+ "\u001b[0;34m >>> ac.goal_test()\u001b[0m\n",
+ "\u001b[0;34m False\u001b[0m\n",
+ "\u001b[0;34m >>> ac.act(expr('Unload(C1, P1, JFK)'))\u001b[0m\n",
+ "\u001b[0;34m >>> ac.goal_test()\u001b[0m\n",
+ "\u001b[0;34m True\u001b[0m\n",
+ "\u001b[0;34m >>>\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mPlanningProblem\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minitial\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(C1, SFO) & At(C2, JFK) & At(P1, SFO) & At(P2, JFK)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mgoals\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(C1, JFK) & At(C2, SFO)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mactions\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Load(c, p, a)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(c, a) & At(p, a)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'In(c, p) & ~At(c, a)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Cargo(c) & Plane(p) & Airport(a)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Unload(c, p, a)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'In(c, p) & At(p, a)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(c, a) & ~In(c, p)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Cargo(c) & Plane(p) & Airport(a)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mAction\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'Fly(p, f, to)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mprecond\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(p, f)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0meffect\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'At(p, to) & ~At(p, f)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Plane(p) & Airport(f) & Airport(to)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdomain\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'Cargo(C1) & Cargo(C2) & Plane(P1) & Plane(P2) & Airport(SFO) & Airport(JFK)'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource air_cargo"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### GraphPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 38,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 9.06 ms, sys: 3 µs, total: 9.06 ms\n",
+ "Wall time: 8.94 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Load(C2, P2, JFK),\n",
+ " Fly(P2, JFK, SFO),\n",
+ " Load(C1, P1, SFO),\n",
+ " Fly(P1, SFO, JFK),\n",
+ " Unload(C1, P1, JFK),\n",
+ " Unload(C2, P2, SFO)]"
+ ]
+ },
+ "execution_count": 38,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time air_cargo_solution = GraphPlan(air_cargo()).execute()\n",
+ "linearize(air_cargo_solution)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### ForwardPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 39,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "838 paths have been expanded and 1288 paths remain in the frontier\n",
+ "CPU times: user 3.56 s, sys: 4 ms, total: 3.57 s\n",
+ "Wall time: 3.56 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Load(C2, P2, JFK),\n",
+ " Fly(P2, JFK, SFO),\n",
+ " Unload(C2, P2, SFO),\n",
+ " Load(C1, P2, SFO),\n",
+ " Fly(P2, SFO, JFK),\n",
+ " Unload(C1, P2, JFK)]"
+ ]
+ },
+ "execution_count": 39,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time air_cargo_solution = uniform_cost_search(ForwardPlan(air_cargo()), display=True).solution()\n",
+ "air_cargo_solution = list(map(lambda action: Expr(action.name, *action.args), air_cargo_solution))\n",
+ "air_cargo_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### ForwardPlan with Ignore Delete Lists Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 40,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "17 paths have been expanded and 54 paths remain in the frontier\n",
+ "CPU times: user 716 ms, sys: 0 ns, total: 716 ms\n",
+ "Wall time: 717 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Load(C2, P2, JFK),\n",
+ " Fly(P2, JFK, SFO),\n",
+ " Unload(C2, P2, SFO),\n",
+ " Load(C1, P2, SFO),\n",
+ " Fly(P2, SFO, JFK),\n",
+ " Unload(C1, P2, JFK)]"
+ ]
+ },
+ "execution_count": 40,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time air_cargo_solution = astar_search(ForwardPlan(air_cargo()), display=True).solution()\n",
+ "air_cargo_solution = list(map(lambda action: Expr(action.name, *action.args), air_cargo_solution))\n",
+ "air_cargo_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### BackwardPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 41,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "506 paths have been expanded and 65 paths remain in the frontier\n",
+ "CPU times: user 970 ms, sys: 0 ns, total: 970 ms\n",
+ "Wall time: 971 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Load(C1, P1, SFO),\n",
+ " Fly(P1, SFO, JFK),\n",
+ " Load(C2, P1, JFK),\n",
+ " Unload(C1, P1, JFK),\n",
+ " Fly(P1, JFK, SFO),\n",
+ " Unload(C2, P1, SFO)]"
+ ]
+ },
+ "execution_count": 41,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time air_cargo_solution = uniform_cost_search(BackwardPlan(air_cargo()), display=True).solution()\n",
+ "air_cargo_solution = list(map(lambda action: Expr(action.name, *action.args), air_cargo_solution))\n",
+ "air_cargo_solution[::-1]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### BackwardPlan with Ignore Delete Lists Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 42,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "23 paths have been expanded and 50 paths remain in the frontier\n",
+ "CPU times: user 1.19 s, sys: 2 µs, total: 1.19 s\n",
+ "Wall time: 1.2 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Load(C2, P2, JFK),\n",
+ " Fly(P2, JFK, SFO),\n",
+ " Unload(C2, P2, SFO),\n",
+ " Load(C1, P2, SFO),\n",
+ " Fly(P2, SFO, JFK),\n",
+ " Unload(C1, P2, JFK)]"
+ ]
+ },
+ "execution_count": 42,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time air_cargo_solution = astar_search(BackwardPlan(air_cargo()), display=True).solution()\n",
+ "air_cargo_solution = list(map(lambda action: Expr(action.name, *action.args), air_cargo_solution))\n",
+ "air_cargo_solution[::-1]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSPlan"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 43,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 6.5 s, sys: 0 ns, total: 6.5 s\n",
+ "Wall time: 6.51 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Load(C1, P1, SFO),\n",
+ " Fly(P1, SFO, JFK),\n",
+ " Load(C2, P1, JFK),\n",
+ " Unload(C1, P1, JFK),\n",
+ " Fly(P1, JFK, SFO),\n",
+ " Unload(C2, P1, SFO)]"
+ ]
+ },
+ "execution_count": 43,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time air_cargo_solution = CSPlan(air_cargo(), 6, arc_heuristic=no_heuristic)\n",
+ "air_cargo_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSPlan with SAT UP Arc Heuristic"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 44,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 13.6 s, sys: 7.98 ms, total: 13.7 s\n",
+ "Wall time: 13.7 s\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "[Load(C1, P1, SFO),\n",
+ " Fly(P1, SFO, JFK),\n",
+ " Load(C2, P1, JFK),\n",
+ " Unload(C1, P1, JFK),\n",
+ " Fly(P1, JFK, SFO),\n",
+ " Unload(C2, P1, SFO)]"
+ ]
+ },
+ "execution_count": 44,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time air_cargo_solution = CSPlan(air_cargo(), 6, arc_heuristic=sat_up)\n",
+ "air_cargo_solution"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## References\n",
+ "\n",
+ "[[1]](#ref-1) Hoffmann, Jörg. 2001. _FF: The fast-forward planning system_.\n",
+ "\n",
+ "[[2]](#ref-2) Kautz, Henry A and Selman, Bart and others. 1992. _Planning as Satisfiability_."
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.7.5rc1"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}
diff --git a/csp.py b/csp.py
index 91a418a3a..6edb48004 100644
--- a/csp.py
+++ b/csp.py
@@ -1,4 +1,4 @@
-"""CSP (Constraint Satisfaction Problems) problems and solvers. (Chapter 6)."""
+"""CSP (Constraint Satisfaction Problems) problems and solvers. (Chapter 6)"""
import string
from operator import eq, neg
@@ -28,9 +28,9 @@ class CSP(search.Problem):
In the textbook and in most mathematical definitions, the
constraints are specified as explicit pairs of allowable values,
but the formulation here is easier to express and more compact for
- most cases. (For example, the n-Queens problem can be represented
- in O(n) space using this notation, instead of O(N^4) for the
- explicit representation.) In terms of describing the CSP as a
+ most cases (for example, the n-Queens problem can be represented
+ in O(n) space using this notation, instead of O(n^4) for the
+ explicit representation). In terms of describing the CSP as a
problem, that's all there is.
However, the class also supports data structures and methods that help you
@@ -88,12 +88,12 @@ def conflict(var2):
def display(self, assignment):
"""Show a human-readable representation of the CSP."""
# Subclasses can print in a prettier way, or display with a GUI
- print('CSP:', self, 'with assignment:', assignment)
+ print(assignment)
# These methods are for the tree and graph-search interface:
def actions(self, state):
- """Return a list of applicable actions: nonconflicting
+ """Return a list of applicable actions: non conflicting
assignments to an unassigned variable."""
if len(state) == len(self.variables):
return []
@@ -160,7 +160,7 @@ def conflicted_vars(self, current):
# ______________________________________________________________________________
-# Constraint Propagation with AC-3
+# Constraint Propagation with AC3
def no_arc_heuristic(csp, queue):
@@ -177,44 +177,55 @@ def AC3(csp, queue=None, removals=None, arc_heuristic=dom_j_up):
queue = {(Xi, Xk) for Xi in csp.variables for Xk in csp.neighbors[Xi]}
csp.support_pruning()
queue = arc_heuristic(csp, queue)
+ checks = 0
while queue:
(Xi, Xj) = queue.pop()
- if revise(csp, Xi, Xj, removals):
+ revised, checks = revise(csp, Xi, Xj, removals, checks)
+ if revised:
if not csp.curr_domains[Xi]:
- return False
+ return False, checks # CSP is inconsistent
for Xk in csp.neighbors[Xi]:
if Xk != Xj:
queue.add((Xk, Xi))
- return True
+ return True, checks # CSP is satisfiable
-def revise(csp, Xi, Xj, removals):
+def revise(csp, Xi, Xj, removals, checks=0):
"""Return true if we remove a value."""
revised = False
for x in csp.curr_domains[Xi][:]:
# If Xi=x conflicts with Xj=y for every possible y, eliminate Xi=x
- if all(not csp.constraints(Xi, x, Xj, y) for y in csp.curr_domains[Xj]):
+ # if all(not csp.constraints(Xi, x, Xj, y) for y in csp.curr_domains[Xj]):
+ conflict = True
+ for y in csp.curr_domains[Xj]:
+ if csp.constraints(Xi, x, Xj, y):
+ conflict = False
+ checks += 1
+ if not conflict:
+ break
+ if conflict:
csp.prune(Xi, x, removals)
revised = True
- return revised
+ return revised, checks
-# Constraint Propagation with AC-3b: an improved version of AC-3 with
-# double-support domain-heuristic
+# Constraint Propagation with AC3b: an improved version
+# of AC3 with double-support domain-heuristic
def AC3b(csp, queue=None, removals=None, arc_heuristic=dom_j_up):
if queue is None:
queue = {(Xi, Xk) for Xi in csp.variables for Xk in csp.neighbors[Xi]}
csp.support_pruning()
queue = arc_heuristic(csp, queue)
+ checks = 0
while queue:
(Xi, Xj) = queue.pop()
# Si_p values are all known to be supported by Xj
# Sj_p values are all known to be supported by Xi
# Dj - Sj_p = Sj_u values are unknown, as yet, to be supported by Xi
- Si_p, Sj_p, Sj_u = partition(csp, Xi, Xj)
+ Si_p, Sj_p, Sj_u, checks = partition(csp, Xi, Xj, checks)
if not Si_p:
- return False
+ return False, checks # CSP is inconsistent
revised = False
for x in set(csp.curr_domains[Xi]) - Si_p:
csp.prune(Xi, x, removals)
@@ -237,6 +248,7 @@ def AC3b(csp, queue=None, removals=None, arc_heuristic=dom_j_up):
if csp.constraints(Xj, vj_p, Xi, vi_p):
conflict = False
Sj_p.add(vj_p)
+ checks += 1
if not conflict:
break
revised = False
@@ -247,10 +259,10 @@ def AC3b(csp, queue=None, removals=None, arc_heuristic=dom_j_up):
for Xk in csp.neighbors[Xj]:
if Xk != Xi:
queue.add((Xk, Xj))
- return True
+ return True, checks # CSP is satisfiable
-def partition(csp, Xi, Xj):
+def partition(csp, Xi, Xj, checks=0):
Si_p = set()
Sj_p = set()
Sj_u = set(csp.curr_domains[Xj])
@@ -265,6 +277,7 @@ def partition(csp, Xi, Xj):
conflict = False
Si_p.add(vi_u)
Sj_p.add(vj_u)
+ checks += 1
if not conflict:
break
# ... and only if no support can be found among the elements in Sj_u, should the elements vj_p in Sj_p be used
@@ -275,12 +288,13 @@ def partition(csp, Xi, Xj):
if csp.constraints(Xi, vi_u, Xj, vj_p):
conflict = False
Si_p.add(vi_u)
+ checks += 1
if not conflict:
break
- return Si_p, Sj_p, Sj_u - Sj_p
+ return Si_p, Sj_p, Sj_u - Sj_p, checks
-# Constraint Propagation with AC-4
+# Constraint Propagation with AC4
def AC4(csp, queue=None, removals=None, arc_heuristic=dom_j_up):
if queue is None:
@@ -290,6 +304,7 @@ def AC4(csp, queue=None, removals=None, arc_heuristic=dom_j_up):
support_counter = Counter()
variable_value_pairs_supported = defaultdict(set)
unsupported_variable_value_pairs = []
+ checks = 0
# construction and initialization of support sets
while queue:
(Xi, Xj) = queue.pop()
@@ -299,13 +314,14 @@ def AC4(csp, queue=None, removals=None, arc_heuristic=dom_j_up):
if csp.constraints(Xi, x, Xj, y):
support_counter[(Xi, x, Xj)] += 1
variable_value_pairs_supported[(Xj, y)].add((Xi, x))
+ checks += 1
if support_counter[(Xi, x, Xj)] == 0:
csp.prune(Xi, x, removals)
revised = True
unsupported_variable_value_pairs.append((Xi, x))
if revised:
if not csp.curr_domains[Xi]:
- return False
+ return False, checks # CSP is inconsistent
# propagation of removed values
while unsupported_variable_value_pairs:
Xj, y = unsupported_variable_value_pairs.pop()
@@ -319,8 +335,8 @@ def AC4(csp, queue=None, removals=None, arc_heuristic=dom_j_up):
unsupported_variable_value_pairs.append((Xi, x))
if revised:
if not csp.curr_domains[Xi]:
- return False
- return True
+ return False, checks # CSP is inconsistent
+ return True, checks # CSP is satisfiable
# ______________________________________________________________________________
@@ -336,17 +352,15 @@ def first_unassigned_variable(assignment, csp):
def mrv(assignment, csp):
"""Minimum-remaining-values heuristic."""
- return argmin_random_tie(
- [v for v in csp.variables if v not in assignment],
- key=lambda var: num_legal_values(csp, var, assignment))
+ return argmin_random_tie([v for v in csp.variables if v not in assignment],
+ key=lambda var: num_legal_values(csp, var, assignment))
def num_legal_values(csp, var, assignment):
if csp.curr_domains:
return len(csp.curr_domains[var])
else:
- return count(csp.nconflicts(var, val, assignment) == 0
- for val in csp.domains[var])
+ return count(csp.nconflicts(var, val, assignment) == 0 for val in csp.domains[var])
# Value ordering
@@ -359,8 +373,7 @@ def unordered_domain_values(var, assignment, csp):
def lcv(var, assignment, csp):
"""Least-constraining-values heuristic."""
- return sorted(csp.choices(var),
- key=lambda val: csp.nconflicts(var, val, assignment))
+ return sorted(csp.choices(var), key=lambda val: csp.nconflicts(var, val, assignment))
# Inference
@@ -443,8 +456,7 @@ def min_conflicts(csp, max_steps=100000):
def min_conflicts_value(csp, var, current):
"""Return the value that will give var the least number of conflicts.
If there is a tie, choose at random."""
- return argmin_random_tie(csp.domains[var],
- key=lambda val: csp.nconflicts(var, val, current))
+ return argmin_random_tie(csp.domains[var], key=lambda val: csp.nconflicts(var, val, current))
# ______________________________________________________________________________
@@ -570,8 +582,7 @@ def MapColoringCSP(colors, neighbors):
specified as a string of the form defined by parse_neighbors."""
if isinstance(neighbors, str):
neighbors = parse_neighbors(neighbors)
- return CSP(list(neighbors.keys()), UniversalDict(colors), neighbors,
- different_values_constraint)
+ return CSP(list(neighbors.keys()), UniversalDict(colors), neighbors, different_values_constraint)
def parse_neighbors(neighbors, variables=None):
@@ -750,7 +761,7 @@ class Sudoku(CSP):
8 . . | 2 . 3 | . . 9
. . 5 | . 1 . | 3 . .
>>> AC3(e); e.display(e.infer_assignment())
- True
+ (True, 6925)
4 8 3 | 9 2 1 | 6 5 7
9 6 7 | 3 4 5 | 8 2 1
2 5 1 | 8 7 6 | 4 9 3
@@ -913,8 +924,7 @@ def display(self, assignment=None):
"""more detailed string representation of CSP"""
if assignment is None:
assignment = {}
- print('CSP(' + str(self.domains) + ', ' + str([str(c) for c in self.constraints]) + ') with assignment: ' +
- str(assignment))
+ print(assignment)
def consistent(self, assignment):
"""assignment is a variable:value dictionary
@@ -1033,36 +1043,52 @@ def GAC(self, orig_domains=None, to_do=None, arc_heuristic=sat_up):
if orig_domains is None:
orig_domains = self.csp.domains
if to_do is None:
- to_do = {(var, const) for const in self.csp.constraints
- for var in const.scope}
+ to_do = {(var, const) for const in self.csp.constraints for var in const.scope}
else:
to_do = to_do.copy()
domains = orig_domains.copy()
to_do = arc_heuristic(to_do)
+ checks = 0
while to_do:
var, const = to_do.pop()
other_vars = [ov for ov in const.scope if ov != var]
+ new_domain = set()
if len(other_vars) == 0:
- new_domain = {val for val in domains[var]
- if const.holds({var: val})}
+ for val in domains[var]:
+ if const.holds({var: val}):
+ new_domain.add(val)
+ checks += 1
+ # new_domain = {val for val in domains[var]
+ # if const.holds({var: val})}
elif len(other_vars) == 1:
other = other_vars[0]
- new_domain = {val for val in domains[var]
- if any(const.holds({var: val, other: other_val})
- for other_val in domains[other])}
- else:
- new_domain = {val for val in domains[var]
- if self.any_holds(domains, const, {var: val}, other_vars)}
+ for val in domains[var]:
+ for other_val in domains[other]:
+ checks += 1
+ if const.holds({var: val, other: other_val}):
+ new_domain.add(val)
+ break
+ # new_domain = {val for val in domains[var]
+ # if any(const.holds({var: val, other: other_val})
+ # for other_val in domains[other])}
+ else: # general case
+ for val in domains[var]:
+ holds, checks = self.any_holds(domains, const, {var: val}, other_vars, checks=checks)
+ if holds:
+ new_domain.add(val)
+ # new_domain = {val for val in domains[var]
+ # if self.any_holds(domains, const, {var: val}, other_vars)}
if new_domain != domains[var]:
domains[var] = new_domain
if not new_domain:
- return False, domains
+ return False, domains, checks
add_to_do = self.new_to_do(var, const).difference(to_do)
to_do |= add_to_do
- return True, domains
+ return True, domains, checks
def new_to_do(self, var, const):
- """returns new elements to be added to to_do after assigning
+ """
+ Returns new elements to be added to to_do after assigning
variable var in constraint const.
"""
return {(nvar, nconst) for nconst in self.csp.var_to_const[var]
@@ -1070,31 +1096,33 @@ def new_to_do(self, var, const):
for nvar in nconst.scope
if nvar != var}
- def any_holds(self, domains, const, env, other_vars, ind=0):
- """returns True if Constraint const holds for an assignment
+ def any_holds(self, domains, const, env, other_vars, ind=0, checks=0):
+ """
+ Returns True if Constraint const holds for an assignment
that extends env with the variables in other_vars[ind:]
env is a dictionary
Warning: this has side effects and changes the elements of env
"""
if ind == len(other_vars):
- return const.holds(env)
+ return const.holds(env), checks + 1
else:
var = other_vars[ind]
for val in domains[var]:
- # env = dict_union(env,{var:val}) # no side effects!
+ # env = dict_union(env, {var:val}) # no side effects
env[var] = val
- holds = self.any_holds(domains, const, env, other_vars, ind + 1)
+ holds, checks = self.any_holds(domains, const, env, other_vars, ind + 1, checks)
if holds:
- return True
- return False
+ return True, checks
+ return False, checks
def domain_splitting(self, domains=None, to_do=None, arc_heuristic=sat_up):
- """return a solution to the current CSP or False if there are no solutions
+ """
+ Return a solution to the current CSP or False if there are no solutions
to_do is the list of arcs to check
"""
if domains is None:
domains = self.csp.domains
- consistency, new_domains = self.GAC(domains, to_do, arc_heuristic)
+ consistency, new_domains, _ = self.GAC(domains, to_do, arc_heuristic)
if not consistency:
return False
elif all(len(new_domains[var]) == 1 for var in domains):
@@ -1120,11 +1148,11 @@ def partition_domain(dom):
class ACSearchSolver(search.Problem):
"""A search problem with arc consistency and domain splitting
- A node is a CSP """
+ A node is a CSP"""
def __init__(self, csp, arc_heuristic=sat_up):
self.cons = ACSolver(csp)
- consistency, self.domains = self.cons.GAC(arc_heuristic=arc_heuristic)
+ consistency, self.domains, _ = self.cons.GAC(arc_heuristic=arc_heuristic)
if not consistency:
raise Exception('CSP is inconsistent')
self.heuristic = arc_heuristic
@@ -1142,7 +1170,7 @@ def actions(self, state):
to_do = self.cons.new_to_do(var, None)
for dom in [dom1, dom2]:
new_domains = extend(state, var, dom)
- consistency, cons_doms = self.cons.GAC(new_domains, to_do, self.heuristic)
+ consistency, cons_doms, _ = self.cons.GAC(new_domains, to_do, self.heuristic)
if consistency:
neighs.append(cons_doms)
return neighs
diff --git a/deep_learning4e.py b/deep_learning4e.py
index 87b33546a..d92a5f3ee 100644
--- a/deep_learning4e.py
+++ b/deep_learning4e.py
@@ -10,7 +10,7 @@
from keras.models import Sequential
from keras.preprocessing import sequence
-from utils4e import (sigmoid, dotproduct, softmax1D, conv1D, GaussianKernel, element_wise_product, vector_add,
+from utils4e import (sigmoid, dot_product, softmax1D, conv1D, GaussianKernel, element_wise_product, vector_add,
random_weights, scalar_vector_product, matrix_multiplication, map_vector, mse_loss)
@@ -107,7 +107,7 @@ def forward(self, inputs):
res = []
# get the output value of each unit
for unit in self.nodes:
- val = self.activation.f(dotproduct(unit.weights, inputs))
+ val = self.activation.f(dot_product(unit.weights, inputs))
unit.val = val
res.append(val)
return res
diff --git a/games.py b/games.py
index d26029fea..cdc24af09 100644
--- a/games.py
+++ b/games.py
@@ -4,9 +4,8 @@
import random
import itertools
import copy
-from utils import argmax, vector_add
+from utils import argmax, vector_add, inf
-inf = float('inf')
GameState = namedtuple('GameState', 'to_move, utility, board, moves')
StochasticGameState = namedtuple('StochasticGameState', 'to_move, utility, board, moves, chance')
diff --git a/games4e.py b/games4e.py
index a79fb5fb3..6bc97c2bb 100644
--- a/games4e.py
+++ b/games4e.py
@@ -4,9 +4,8 @@
import random
import itertools
import copy
-from utils import argmax, vector_add, MCT_Node, ucb
+from utils4e import argmax, vector_add, MCT_Node, ucb, inf
-inf = float('inf')
GameState = namedtuple('GameState', 'to_move, utility, board, moves')
StochasticGameState = namedtuple('StochasticGameState', 'to_move, utility, board, moves, chance')
@@ -187,8 +186,8 @@ def select(n):
def expand(n):
"""expand the leaf node by adding all its children states"""
if not n.children and not game.terminal_test(n.state):
- n.children = {MCT_Node(state=game.result(n.state, action), parent=n): action for action in
- game.actions(n.state)}
+ n.children = {MCT_Node(state=game.result(n.state, action), parent=n): action
+ for action in game.actions(n.state)}
return select(n)
def simulate(game, state):
diff --git a/improving_sat_algorithms.ipynb b/improving_sat_algorithms.ipynb
new file mode 100644
index 000000000..d461e99c4
--- /dev/null
+++ b/improving_sat_algorithms.ipynb
@@ -0,0 +1,2539 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "# Propositional Logic\n",
+ "---\n",
+ "# Improving Boolean Satisfiability Algorithms\n",
+ "\n",
+ "## Introduction\n",
+ "A propositional formula $\\Phi$ in *Conjunctive Normal Form* (CNF) is a conjunction of clauses $\\omega_j$, with $j \\in \\{1,...,m\\}$. Each clause being a disjunction of literals and each literal being either a positive ($x_i$) or a negative ($\\lnot{x_i}$) propositional variable, with $i \\in \\{1,...,n\\}$. By denoting with $[\\lnot]$ the possible presence of $\\lnot$, we can formally define $\\Phi$ as:\n",
+ "\n",
+ "$$\\bigwedge_{j = 1,...,m}\\bigg(\\bigvee_{i \\in \\omega_j} [\\lnot] x_i\\bigg)$$\n",
+ "\n",
+ "The ***Boolean Satisfiability Problem*** (SAT) consists in determining whether there exists a truth assignment in $\\{0, 1\\}$ (or equivalently in $\\{True,False\\}$) for the variables in $\\Phi$."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from logic import *"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## DPLL with Branching Heuristics\n",
+ "The ***Davis-Putnam-Logemann-Loveland*** (DPLL) algorithm is a *complete* (will answer SAT if a solution exists) and *sound* (it will not answer SAT for an unsatisfiable formula) procedue that combines *backtracking search* and *deduction* to decide satisfiability of propositional logic formula in CNF. At each search step a variable and a propositional value are selected for branching purposes. With each branching step, two values can be assigned to a variable, either 0 or 1. Branching corresponds to assigning the chosen value to the chosen variable. Afterwards, the logical consequences of each branching step are evaluated. Each time an unsatisfied clause (ie a *conflict*) is identified, backtracking is executed. Backtracking corresponds to undoing branching steps until an unflipped branch is reached. When both values have been assigned to the selected variable at a branching step, backtracking will undo this branching step. If for the first branching step both values have been considered, and backtracking undoes this first branching step, then the CNF formula can be declared unsatisfiable. This kind of backtracking is called *chronological backtracking*.\n",
+ "\n",
+ "Essentially, `DPLL` is a backtracking depth-first search through partial truth assignments which uses a *splitting rule* to replaces the original problem with two smaller subproblems, whereas the original Davis-Putnam procedure uses a variable elimination rule which replaces the original problem with one larger subproblem. Over the years, many heuristics have been proposed in choosing the splitting variable (which variable should be assigned a truth value next).\n",
+ "\n",
+ "Search algorithms that are based on a predetermined order of search are called static algorithms, whereas the ones that select them at the runtime are called dynamic. The first SAT search algorithm, the Davis-Putnam procedure is a static algorithm. Static search algorithms are usually very slow in practice and for this reason perform worse than dynamic search algorithms. However, dynamic search algorithms are much harder to design, since they require a heuristic for predetermining the order of search. The fundamental element of a heuristic is a branching strategy for selecting the next branching literal. This must not require a lot of time to compute and yet it must provide a powerful insight into the problem instance.\n",
+ "\n",
+ "Two basic heuristics are applied to this algorithm with the potential of cutting the search space in half. These are the *pure literal rule* and the *unit clause rule*.\n",
+ "- the *pure literal* rule is applied whenever a variable appears with a single polarity in all the unsatisfied clauses. In this case, assigning a truth value to the variable so that all the involved clauses are satisfied is highly effective in the search;\n",
+ "- if some variable occurs in the current formula in a clause of length 1 then the *unit clause* rule is applied. Here, the literal is selected and a truth value so the respective clause is satisfied is assigned. The iterative application of the unit rule is commonly reffered to as *Boolean Constraint Propagation* (BCP)."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mdpll_satisfiable\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbranching_heuristic\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mno_branching_heuristic\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"Check satisfiability of a propositional sentence.\u001b[0m\n",
+ "\u001b[0;34m This differs from the book code in two ways: (1) it returns a model\u001b[0m\n",
+ "\u001b[0;34m rather than True when it succeeds; this is more useful. (2) The\u001b[0m\n",
+ "\u001b[0;34m function find_pure_symbol is passed a list of unknown clauses, rather\u001b[0m\n",
+ "\u001b[0;34m than a list of all clauses and the model; this is more efficient.\u001b[0m\n",
+ "\u001b[0;34m >>> dpll_satisfiable(A |'<=>'| B) == {A: True, B: True}\u001b[0m\n",
+ "\u001b[0;34m True\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mdpll\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mto_cnf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mprop_symbols\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbranching_heuristic\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource dpll_satisfiable"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mdpll\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbranching_heuristic\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mno_branching_heuristic\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"See if the clauses are true in a partial model.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0munknown_clauses\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;31m# clauses with an unknown truth value\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mval\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpl_true\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mval\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mval\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0munknown_clauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0munknown_clauses\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfind_pure_symbol\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0munknown_clauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mdpll\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremove_all\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mextend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbranching_heuristic\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfind_unit_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mdpll\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremove_all\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mextend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbranching_heuristic\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbranching_heuristic\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0munknown_clauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mdpll\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremove_all\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mextend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbranching_heuristic\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mor\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdpll\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremove_all\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mextend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbranching_heuristic\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource dpll"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Each of these branching heuristics was applied only after the *pure literal* and the *unit clause* heuristic failed in selecting a splitting variable."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### MOMs"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "MOMs heuristics are simple, efficient and easy to implement. The goal of these heuristics is to prefer the literal having ***Maximum number of Occurences in the Minimum length clauses***. Intuitively, the literals belonging to the minimum length clauses are the most constrained literals in the formula. Branching on them will maximize the effect of BCP and the likelihood of hitting a dead end early in the search tree (for unsatisfiable problems). Conversely, in the case of satisfiable formulas, branching on a highly constrained variable early in the tree will also increase the likelihood of a correct assignment of the remained open literals.\n",
+ "The MOMs heuristics main disadvatage is that their effectiveness highly depends on the problem instance. It is easy to see that the ideal setting for these heuristics is considering the unsatisfied binary clauses."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mmin_clauses\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mmin_len\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mc\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdefault\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0mc\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mmin_len\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mmin_len\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource min_clauses"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mmoms\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m MOMS (Maximum Occurrence in clauses of Minimum Size) heuristic\u001b[0m\n",
+ "\u001b[0;34m Returns the literal with the most occurrences in all clauses of minimum size\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmin_clauses\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mprop_symbols\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource moms"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Over the years, many types of MOMs heuristics have been proposed.\n",
+ "\n",
+ "***MOMSf*** choose the variable $x$ with a maximize the function:\n",
+ "\n",
+ "$$[f(x) + f(\\lnot{x})] * 2^k + f(x) * f(\\lnot{x})$$\n",
+ "\n",
+ "where $f(x)$ is the number of occurrences of $x$ in the smallest unknown clauses, k is a parameter."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mmomsf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m MOMS alternative heuristic\u001b[0m\n",
+ "\u001b[0;34m If f(x) the number of occurrences of the variable x in clauses with minimum size,\u001b[0m\n",
+ "\u001b[0;34m we choose the variable maximizing [f(x) + f(-x)] * 2^k + f(x) * f(-x)\u001b[0m\n",
+ "\u001b[0;34m Returns x if f(x) >= f(-x) otherwise -x\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmin_clauses\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mpow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m>=\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource momsf"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "***Freeman’s POSIT*** [[1]](#cite-freeman1995improvements) version counts both the number of positive $x$ and negative $\\lnot{x}$ occurrences of a given variable $x$."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mposit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Freeman's POSIT version of MOMs\u001b[0m\n",
+ "\u001b[0;34m Counts the positive x and negative x for each variable x in clauses with minimum size\u001b[0m\n",
+ "\u001b[0;34m Returns x if f(x) >= f(-x) otherwise -x\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmin_clauses\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m>=\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource posit"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "***Zabih and McAllester’s*** [[2]](#cite-zabih1988rearrangement) version of the heuristic counts the negative occurrences $\\lnot{x}$ of each given variable $x$."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mzm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Zabih and McAllester's version of MOMs\u001b[0m\n",
+ "\u001b[0;34m Counts the negative occurrences only of each variable x in clauses with minimum size\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmin_clauses\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0ml\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mop\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'~'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource zm"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### DLIS & DLCS"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Literal count heuristics count the number of unresolved clauses in which a given variable $x$ appears as a positive literal, $C_P$ , and as negative literal, $C_N$. These two numbers an either be onsidered individually or ombined. \n",
+ "\n",
+ "***Dynamic Largest Individual Sum*** heuristic considers the values $C_P$ and $C_N$ separately: select the variable with the largest individual value and assign to it value true if $C_P \\geq C_N$, value false otherwise."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mdlis\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m DLIS (Dynamic Largest Individual Sum) heuristic\u001b[0m\n",
+ "\u001b[0;34m Choose the variable and value that satisfies the maximum number of unsatisfied clauses\u001b[0m\n",
+ "\u001b[0;34m Like DLCS but we only consider the literal (thus Cp and Cn are individual)\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mclauses\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m>=\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource dlis"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "***Dynamic Largest Combined Sum*** considers the values $C_P$ and $C_N$ combined: select the variable with the largest sum $C_P + C_N$ and assign to it value true if $C_P \\geq C_N$, value false otherwise."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mdlcs\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m DLCS (Dynamic Largest Combined Sum) heuristic\u001b[0m\n",
+ "\u001b[0;34m Cp the number of clauses containing literal x\u001b[0m\n",
+ "\u001b[0;34m Cn the number of clauses containing literal -x\u001b[0m\n",
+ "\u001b[0;34m Here we select the variable maximizing Cp + Cn\u001b[0m\n",
+ "\u001b[0;34m Returns x if Cp >= Cn otherwise -x\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mclauses\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m>=\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource dlcs"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### JW & JW2"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Two branching heuristics were proposed by ***Jeroslow and Wang*** in [[3]](#cite-jeroslow1990solving).\n",
+ "\n",
+ "The *one-sided Jeroslow and Wang*’s heuristic compute:\n",
+ "\n",
+ "$$J(l) = \\sum_{l \\in \\omega \\land \\omega \\in \\phi} 2^{-|\\omega|}$$\n",
+ "\n",
+ "and selects the assignment that satisfies the literal with the largest value $J(l)$."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mjw\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Jeroslow-Wang heuristic\u001b[0m\n",
+ "\u001b[0;34m For each literal compute J(l) = \\sum{l in clause c} 2^{-|c|}\u001b[0m\n",
+ "\u001b[0;34m Return the literal maximizing J\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mprop_symbols\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mpow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m-\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource jw"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "The *two-sided Jeroslow and Wang*’s heuristic identifies the variable $x$ with the largest sum $J(x) + J(\\lnot{x})$, and assigns to $x$ value true, if $J(x) \\geq J(\\lnot{x})$, and value false otherwise."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mjw2\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m Two Sided Jeroslow-Wang heuristic\u001b[0m\n",
+ "\u001b[0;34m Compute J(l) also counts the negation of l = J(x) + J(-x)\u001b[0m\n",
+ "\u001b[0;34m Returns x if J(x) >= J(-x) otherwise -x\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mpow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m-\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m>=\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource jw2"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## CDCL with 1UIP Learning Scheme, 2WL Lazy Data Structure, VSIDS Branching Heuristic & Restarts\n",
+ "\n",
+ "The ***Conflict-Driven Clause Learning*** (CDCL) solver is an evolution of the *DPLL* algorithm that involves a number of additional key techniques:\n",
+ "\n",
+ "- non-chronological backtracking or *backjumping*;\n",
+ "- *learning* new *clauses* from conflicts during search by exploiting its structure;\n",
+ "- using *lazy data structures* for storing clauses;\n",
+ "- *branching heuristics* with low computational overhead and which receive feedback from search;\n",
+ "- periodically *restarting* search.\n",
+ "\n",
+ "The first difference between a DPLL solver and a CDCL solver is the introduction of the *non-chronological backtracking* or *backjumping* when a conflict is identified. This requires an iterative implementation of the algorithm because only if the backtrack stack is managed explicitly it is possible to backtrack more than one level."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 4,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mcdcl_satisfiable\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvsids_decay\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.95\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrestart_strategy\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mno_restart\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m >>> cdcl_satisfiable(A |'<=>'| B) == {A: True, B: True}\u001b[0m\n",
+ "\u001b[0;34m True\u001b[0m\n",
+ "\u001b[0;34m \"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mTwoWLClauseDatabase\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mto_cnf\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msymbols\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mprop_symbols\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ms\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mCounter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mG\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mDiGraph\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdl\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflicts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrestarts\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msum_lbd\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue_lbd\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0munit_propagation\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mconflict\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mdl\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflicts\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlearn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlbd\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mconflict_analysis\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue_lbd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlbd\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msum_lbd\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mlbd\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbackjump\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlearn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mupdate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlearn\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0msymbol\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m*=\u001b[0m \u001b[0mvsids_decay\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mrestart_strategy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconflicts\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrestarts\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue_lbd\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msum_lbd\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbackjump\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mqueue_lbd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mclear\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mrestarts\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdl\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0massign_decision_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource cdcl_satisfiable"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Clause Learning with 1UIP Scheme"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "The second important difference between a DPLL solver and a CDCL solver is that the information about a conflict is reused by learning: if a conflicting clause is found, the solver derive a new clause from the conflict and add it to the clauses database.\n",
+ "\n",
+ "Whenever a conflict is identified due to unit propagation, a conflict analysis procedure is invoked. As a result, one or more new clauses are learnt, and a backtracking decision level is computed. The conflict analysis procedure analyzes the structure of unit propagation and decides which literals to include in the learnt clause. The decision levels associated with assigned variables define a partial order of the variables. Starting from a given unsatisfied clause (represented in the implication graph with vertex $\\kappa$), the conflict analysis procedure visits variables implied at the most recent decision level (ie the current largest decision level), identifies the antecedents of visited variables, and keeps from the antecedents the literals assigned at decision levels less than the most recent decision level. The clause learning procedure used in the CDCL can be defined by a sequence of selective resolution operations, that at each step yields a new temporary clause. This process is repeated until the most recent decision variable is visited.\n",
+ "\n",
+ "The structure of implied assignments induced by unit propagation is a key aspect of the clause learning procedure. Moreover, the idea of exploiting the structure induced by unit propagation was further exploited with ***Unit Implication Points*** (UIPs). A UIP is a *dominator* in the implication graph and represents an alternative decision assignment at the current decision level that results in the same conflict. The main motivation for identifying UIPs is to reduce the size of learnt clauses. Clause learning could potentially stop at any UIP, being quite straightforward to conclude that the set of literals of a clause learnt at the first UIP has clear advantages. Considering the largest decision level of the literals of the clause learnt at each UIP, the clause learnt at the first UIP is guaranteed to contain the smallest one. This guarantees the highest backtrack jump in the search tree."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mconflict_analysis\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict_clause\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mG\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'K'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'antecedent'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mp\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpred\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'K'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnode\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mnode\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnodes\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;34m'K'\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnodes\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnode\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'dl'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mdl\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0min_degree\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnode\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mfirst_uip\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mimmediate_dominators\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mP\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'K'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mremove_node\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'K'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict_side\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdescendants\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfirst_uip\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mprop_symbols\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconflict_clause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mintersection\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconflict_side\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mantecedent\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mG\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'antecedent'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mp\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpred\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict_clause\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpl_binary_resolution\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconflict_clause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mantecedent\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# the literal block distance is calculated by taking the decision levels from variables of all\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# literals in the clause, and counting how many different decision levels were in this set\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mlbd\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnodes\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'dl'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mprop_symbols\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconflict_clause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlbd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcount\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mfirst_uip\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mprop_symbols\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconflict_clause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;36m0\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlbd\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mheapq\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnlargest\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlbd\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mconflict_clause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlbd\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource conflict_analysis"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mpl_binary_resolution\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mci\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mdi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mci\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mdj\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mdi\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m~\u001b[0m\u001b[0mdj\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;34m~\u001b[0m\u001b[0mdi\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mdj\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mpl_binary_resolution\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'|'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremove_all\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdi\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mci\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'|'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mremove_all\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdj\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0massociate\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'|'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0munique\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mci\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcj\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource pl_binary_resolution"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mbackjump\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mdelete\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mnode\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mnode\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnodes\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnodes\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnode\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'dl'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mremove_nodes_from\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdelete\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mnode\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdelete\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdel\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnode\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msymbols\u001b[0m \u001b[0;34m|=\u001b[0m \u001b[0mdelete\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource backjump"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### 2WL Lazy Data Structure"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Implementation issues for SAT solvers include the design of suitable data structures for storing clauses. The implemented data structures dictate the way BCP are implemented and have a significant impact on the run time performance of the SAT solver. Recent state-of-the-art SAT solvers are characterized by using very efficient data structures, intended to reduce the CPU time required per each node in the search tree. Conversely, traditional SAT data structures are accurate, meaning that is possible to know exactly the value of each literal in the clause. Examples of the most recent SAT data structures, which are not accurate and therefore are called lazy, include the watched literals used in Chaff .\n",
+ "\n",
+ "The more recent Chaff SAT solver [[4]](#cite-moskewicz2001chaff) proposed a new data structure, the ***2 Watched Literals*** (2WL), in which two references are associated with each clause. There is no order relation between the two references, allowing the references to move in any direction. The lack of order between the two references has the key advantage that no literal references need to be updated when backtracking takes place. In contrast, unit or unsatisfied clauses are identified only after traversing all the clauses’ literals; a clear drawback. The two watched literal pointers are undifferentiated as there is no order relation. Again, each time one literal pointed by one of these pointers is assigned, the pointer has to move inwards. These pointers may move in both directions. This causes the whole clause to be traversed when the clause becomes unit. In addition, no references have to be kept to the just assigned literals, since pointers do not move when backtracking."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0munit_propagation\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mcheck\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mmodel\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0m_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mw1\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_neg_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mw1\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw1\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_pos_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mw1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0m_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mw2\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_neg_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mw2\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw2\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_pos_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mw2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0munit_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mwatching\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mwatching\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd_node\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mval\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mp\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd_edges_from\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprop_symbols\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mitertools\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcycle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mantecedent\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mremove\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mp\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mconflict_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd_edges_from\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mzip\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mprop_symbols\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mitertools\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcycle\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'K'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mantecedent\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbcp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mfilter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcheck\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_clauses\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# we need only visit each clause when one of its two watched literals is assigned to 0 because, until\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# this happens, we can guarantee that there cannot be more than n-2 literals in the clause assigned to 0\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mfirst_watched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpl_true\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msecond_watched\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpl_true\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfirst_watched\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0munit_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbcp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melif\u001b[0m \u001b[0mfirst_watched\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mFalse\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0msecond_watched\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mupdate_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbcp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if the only literal with a non-zero value is the other watched literal then\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0msecond_watched\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;31m# if it is free, then the clause is a unit clause\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0munit_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbcp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;31m# else (it is False) the clause is a conflict clause\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melif\u001b[0m \u001b[0msecond_watched\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mFalse\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mfirst_watched\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mupdate_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbcp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if the only literal with a non-zero value is the other watched literal then\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfirst_watched\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;31m# if it is free, then the clause is a unit clause\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0munit_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclauses\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbcp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mbreak\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;31m# else (it is False) the clause is a conflict clause\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mconflict_clause\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mbcp\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource unit_propagation"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mclass\u001b[0m \u001b[0mTwoWLClauseDatabase\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__twl\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m{\u001b[0m\u001b[0;34m}\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdefaultdict\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mc\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mclauses\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mc\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mget_clauses\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__twl\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeys\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mset_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnew_watching\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__twl\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnew_watching\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mset_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnew_watching\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__twl\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnew_watching\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__twl\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__twl\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mget_pos_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0ml\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mget_neg_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0ml\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__twl\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__assign_watching_literals\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp1\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mp1\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mw1\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mw2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mp2\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mremove\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp1\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdel\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__twl\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdiscard\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mp1\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdiscard\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mw1\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mw2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdiscard\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mp2\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw2\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdiscard\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mupdate_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if a non-zero literal different from the other watched literal is found\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mfound\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnew_watching\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__find_new_watching_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfound\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;31m# then it will replace the watched literal\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mremove\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mp\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mremove\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnew_watching\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_watching\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mp\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0mupdate_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if a non-zero literal different from the other watched literal is found\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mfound\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnew_watching\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__find_new_watching_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_second_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mfound\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0;31m# then it will replace the watched literal\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mremove\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mp\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mremove\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_first_watched\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnew_watching\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mw\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minspect_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnew_watching\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mp\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__watch_list\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mw\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__find_new_watching_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mother_watched\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if a non-zero literal different from the other watched literal is found\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0ml\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0mother_watched\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mpl_true\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# then it is returned\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0ml\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__assign_watching_literals\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mmodel\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclause\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mnext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mpl_true\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnext\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0ml\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdisjuncts\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mclause\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mpl_true\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0ml\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource TwoWLClauseDatabase"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### VSIDS Branching Heuristic"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "The early branching heuristics made use of all the information available from the data structures, namely the number of satisfied, unsatisfied and unassigned literals. These heuristics are updated during the search and also take into account the clauses that are learnt. \n",
+ "\n",
+ "More recently, a different kind of variable selection heuristic, referred to as ***Variable State Independent Decaying Sum*** (VSIDS), has been proposed by Chaff authors in [[4]](#cite-moskewicz2001chaff). One of the reasons for proposing this new heuristic was the introduction of lazy data structures, where the knowledge of the dynamic size of a clause is not accurate. Hence, the heuristics described above cannot be used. VSIDS selects the literal that appears most frequently over all the clauses, which means that one counter is required for each one of the literals. Initially, all counters are set to zero. During the search, the metrics only have to be updated when a new recorded clause is created. More than to develop an accurate heuristic, the motivation has been to design a fast (but dynamically adapting) heuristic. In fact, one of the key properties of this strategy is the very low overhead, due to being independent of the variable state."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0massign_decision_literal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mP\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msymbols\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkey\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0msymbol\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0msymbol\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mvalue\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m>=\u001b[0m \u001b[0mscores\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m~\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0msymbols\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mremove\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mG\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd_node\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mP\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mval\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mvalue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdl\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mdl\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource assign_decision_literal"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Restarts"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Solving NP-complete problems, such as SAT, naturally leads to heavy-tailed run times. To deal with this, SAT solvers frequently restart their search to avoid the runs that take disproportionately longer. What restarting here means is that the solver unsets all variables and starts the search using different variable assignment order.\n",
+ "\n",
+ "While at first glance it might seem that restarts should be rare and become rarer as the solving has been going on for longer, so that the SAT solver can actually finish solving the problem, the trend has been towards more aggressive (frequent) restarts.\n",
+ "\n",
+ "The reason why frequent restarts help solve problems faster is that while the solver does forget all current variable assignments, it does keep some information, specifically it keeps learnt clauses, effectively sampling the search space, and it keeps the last assigned truth value of each variable, assigning them the same value the next time they are picked to be assigned."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### Luby"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In this strategy, the number of conflicts between 2 restarts is based on the *Luby* sequence. The *Luby* restart sequence is interesting in that it was proven to be optimal restart strategy for randomized search algorithms where the runs do not share information. While this is not true for SAT solving, as shown in [[5]](cite-haim2014towards) and [[6]](cite-huang2007effect), *Luby* restarts have been quite successful anyway.\n",
+ "\n",
+ "The exact description of *Luby* restarts is that the $ith$ restart happens after $u \\cdot Luby(i)$ conflicts, where $u$ is a constant and $Luby(i)$ is defined as:\n",
+ "\n",
+ "$$Luby(i) = \\begin{cases} \n",
+ " 2^{k-1} & i = 2^k - 1 \\\\\n",
+ " Luby(i - 2^{k-1} + 1) & 2^{k-1} \\leq i < 2^k - 1\n",
+ " \\end{cases}\n",
+ "$$\n",
+ "\n",
+ "A less exact but more intuitive description of the *Luby* sequence is that all numbers in it are powers of two, and after a number is seen for the second time, the next number is twice as big. The following are the first 16 numbers in the sequence:\n",
+ "\n",
+ "$$ (1,1,2,1,1,2,4,1,1,2,1,1,2,4,8,1,...) $$\n",
+ "\n",
+ "From the above, we can see that this restart strategy tends towards frequent restarts, but some runs are kept running for much longer, and there is no upper limit on the longest possible time between two restarts."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mluby\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconflicts\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrestarts\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue_lbd\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msum_lbd\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0munit\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m512\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# in the state-of-art tested with unit value 1, 2, 4, 6, 8, 12, 16, 32, 64, 128, 256 and 512\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m_luby\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mk\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mwhile\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mi\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m \u001b[0;34m<<\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0;36m1\u001b[0m \u001b[0;34m<<\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mk\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melif\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m \u001b[0;34m<<\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mk\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m<=\u001b[0m \u001b[0mi\u001b[0m \u001b[0;34m<\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m \u001b[0;34m<<\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0m_luby\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mi\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m \u001b[0;34m<<\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mk\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mk\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0munit\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0m_luby\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrestarts\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mqueue_lbd\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource luby"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### Glucose"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Glucose restarts were popularized by the *Glucose* solver, and it is an extremely aggressive, dynamic restart strategy. The idea behind it and described in [[7]](cite-audemard2012refining) is that instead of waiting for a fixed amount of conflicts, we restart when the last couple of learnt clauses are, on average, bad.\n",
+ "\n",
+ "A bit more precisely, if there were at least $X$ conflicts (and thus $X$ learnt clauses) since the last restart, and the average *Literal Block Distance* (LBD) (a criterion to evaluate the quality of learnt clauses as shown in [[8]](#cite-audemard2009predicting) of the last $X$ learnt clauses was at least $K$ times higher than the average LBD of all learnt clauses, it is time for another restart. Parameters $X$ and $K$ can be tweaked to achieve different restart frequency, and they are usually kept quite small."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mglucose\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mconflicts\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrestarts\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mqueue_lbd\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msum_lbd\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mk\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.7\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# in the state-of-art tested with (x, k) as (50, 0.8) and (100, 0.7)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# if there were at least x conflicts since the last restart, and then the average LBD of the last\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# x learnt clauses was at least k times higher than the average LBD of all learnt clauses\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mqueue_lbd\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>=\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0msum\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mqueue_lbd\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mqueue_lbd\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mk\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0msum_lbd\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mconflicts\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource glucose"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "pycharm": {}
+ },
+ "source": [
+ "## Experimental Results"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from csp import *"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Australia"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSP"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "australia_csp = MapColoringCSP(list('RGB'), \"\"\"SA: WA NT Q NSW V; NT: WA Q; NSW: Q V; T: \"\"\")"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 154 µs, sys: 37 µs, total: 191 µs\n",
+ "Wall time: 194 µs\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b with DOM J UP needs 72 consistency-checks'"
+ ]
+ },
+ "execution_count": 24,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, checks = AC3b(australia_csp, arc_heuristic=dom_j_up)\n",
+ "f'AC3b with DOM J UP needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 25,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 263 µs, sys: 0 ns, total: 263 µs\n",
+ "Wall time: 268 µs\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "{'Q': 'R', 'SA': 'G', 'NSW': 'B', 'NT': 'B', 'V': 'R', 'WA': 'R'}"
+ ]
+ },
+ "execution_count": 25,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time backtracking_search(australia_csp, select_unassigned_variable=mrv, inference=forward_checking)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### SAT"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "australia_sat = MapColoringSAT(list('RGB'), \"\"\"SA: WA NT Q NSW V; NT: WA Q; NSW: Q V; T: \"\"\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "##### DPLL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 3,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 43.3 ms, sys: 0 ns, total: 43.3 ms\n",
+ "Wall time: 41.5 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(australia_sat, branching_heuristic=no_branching_heuristic)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 36.4 ms, sys: 0 ns, total: 36.4 ms\n",
+ "Wall time: 35.3 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(australia_sat, branching_heuristic=moms)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 29,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 36.1 ms, sys: 3.9 ms, total: 40 ms\n",
+ "Wall time: 39.2 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(australia_sat, branching_heuristic=momsf)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 30,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 45.2 ms, sys: 0 ns, total: 45.2 ms\n",
+ "Wall time: 44.2 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(australia_sat, branching_heuristic=posit)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 31,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 31.2 ms, sys: 0 ns, total: 31.2 ms\n",
+ "Wall time: 30.5 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(australia_sat, branching_heuristic=zm)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 32,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 57 ms, sys: 0 ns, total: 57 ms\n",
+ "Wall time: 55.9 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(australia_sat, branching_heuristic=dlis)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 33,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 51.8 ms, sys: 0 ns, total: 51.8 ms\n",
+ "Wall time: 50.7 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(australia_sat, branching_heuristic=dlcs)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 34,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 40.6 ms, sys: 0 ns, total: 40.6 ms\n",
+ "Wall time: 39.3 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(australia_sat, branching_heuristic=jw)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 35,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 43.2 ms, sys: 1.81 ms, total: 45.1 ms\n",
+ "Wall time: 43.9 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(australia_sat, branching_heuristic=jw2)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "##### CDCL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 36,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 32.9 ms, sys: 16 µs, total: 33 ms\n",
+ "Wall time: 31.6 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = cdcl_satisfiable(australia_sat)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 37,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "{NSW_B, NT_B, Q_G, SA_R, V_G, WA_G}"
+ ]
+ },
+ "execution_count": 37,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "{var for var, val in model.items() if val}"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### France"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSP"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 38,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "france_csp = MapColoringCSP(list('RGBY'),\n",
+ " \"\"\"AL: LO FC; AQ: MP LI PC; AU: LI CE BO RA LR MP; BO: CE IF CA FC RA\n",
+ " AU; BR: NB PL; CA: IF PI LO FC BO; CE: PL NB NH IF BO AU LI PC; FC: BO\n",
+ " CA LO AL RA; IF: NH PI CA BO CE; LI: PC CE AU MP AQ; LO: CA AL FC; LR:\n",
+ " MP AU RA PA; MP: AQ LI AU LR; NB: NH CE PL BR; NH: PI IF CE NB; NO:\n",
+ " PI; PA: LR RA; PC: PL CE LI AQ; PI: NH NO CA IF; PL: BR NB CE PC; RA:\n",
+ " AU BO FC PA LR\"\"\")"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 39,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 599 µs, sys: 112 µs, total: 711 µs\n",
+ "Wall time: 716 µs\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b with DOM J UP needs 516 consistency-checks'"
+ ]
+ },
+ "execution_count": 39,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, checks = AC3b(france_csp, arc_heuristic=dom_j_up)\n",
+ "f'AC3b with DOM J UP needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 40,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 560 µs, sys: 0 ns, total: 560 µs\n",
+ "Wall time: 563 µs\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "{'NH': 'R',\n",
+ " 'NB': 'G',\n",
+ " 'CE': 'B',\n",
+ " 'PL': 'R',\n",
+ " 'BR': 'B',\n",
+ " 'IF': 'G',\n",
+ " 'PI': 'B',\n",
+ " 'BO': 'R',\n",
+ " 'CA': 'Y',\n",
+ " 'FC': 'G',\n",
+ " 'LO': 'R',\n",
+ " 'PC': 'G',\n",
+ " 'AU': 'G',\n",
+ " 'AL': 'B',\n",
+ " 'RA': 'B',\n",
+ " 'LR': 'R',\n",
+ " 'LI': 'R',\n",
+ " 'AQ': 'B',\n",
+ " 'MP': 'Y',\n",
+ " 'PA': 'G',\n",
+ " 'NO': 'R'}"
+ ]
+ },
+ "execution_count": 40,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time backtracking_search(france_csp, select_unassigned_variable=mrv, inference=forward_checking)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### SAT"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 41,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "france_sat = MapColoringSAT(list('RGBY'),\n",
+ " \"\"\"AL: LO FC; AQ: MP LI PC; AU: LI CE BO RA LR MP; BO: CE IF CA FC RA\n",
+ " AU; BR: NB PL; CA: IF PI LO FC BO; CE: PL NB NH IF BO AU LI PC; FC: BO\n",
+ " CA LO AL RA; IF: NH PI CA BO CE; LI: PC CE AU MP AQ; LO: CA AL FC; LR:\n",
+ " MP AU RA PA; MP: AQ LI AU LR; NB: NH CE PL BR; NH: PI IF CE NB; NO:\n",
+ " PI; PA: LR RA; PC: PL CE LI AQ; PI: NH NO CA IF; PL: BR NB CE PC; RA:\n",
+ " AU BO FC PA LR\"\"\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "##### DPLL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 42,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 3.32 s, sys: 0 ns, total: 3.32 s\n",
+ "Wall time: 3.32 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(france_sat, branching_heuristic=no_branching_heuristic)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 43,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 3.17 s, sys: 390 µs, total: 3.17 s\n",
+ "Wall time: 3.17 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(france_sat, branching_heuristic=moms)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 44,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 3.49 s, sys: 0 ns, total: 3.49 s\n",
+ "Wall time: 3.49 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(france_sat, branching_heuristic=momsf)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 45,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 3.5 s, sys: 0 ns, total: 3.5 s\n",
+ "Wall time: 3.5 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(france_sat, branching_heuristic=posit)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 46,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 3 s, sys: 2.6 ms, total: 3.01 s\n",
+ "Wall time: 3.01 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(france_sat, branching_heuristic=zm)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 47,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 12.5 s, sys: 11.4 ms, total: 12.5 s\n",
+ "Wall time: 12.5 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(france_sat, branching_heuristic=dlis)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 48,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 3.41 s, sys: 0 ns, total: 3.41 s\n",
+ "Wall time: 3.41 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(france_sat, branching_heuristic=dlcs)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 49,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 2.92 s, sys: 3.89 ms, total: 2.92 s\n",
+ "Wall time: 2.92 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(france_sat, branching_heuristic=jw)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 50,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 3.71 s, sys: 0 ns, total: 3.71 s\n",
+ "Wall time: 3.73 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(france_sat, branching_heuristic=jw2)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "##### CDCL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 51,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 159 ms, sys: 3.94 ms, total: 163 ms\n",
+ "Wall time: 162 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = cdcl_satisfiable(france_sat)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 52,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "{AL_G,\n",
+ " AQ_G,\n",
+ " AU_R,\n",
+ " BO_G,\n",
+ " BR_Y,\n",
+ " CA_R,\n",
+ " CE_B,\n",
+ " FC_B,\n",
+ " IF_Y,\n",
+ " LI_Y,\n",
+ " LO_Y,\n",
+ " LR_G,\n",
+ " MP_B,\n",
+ " NB_R,\n",
+ " NH_G,\n",
+ " NO_Y,\n",
+ " PA_B,\n",
+ " PC_R,\n",
+ " PI_B,\n",
+ " PL_G,\n",
+ " RA_Y}"
+ ]
+ },
+ "execution_count": 52,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "{var for var, val in model.items() if val}"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### USA"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSP"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 53,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "usa_csp = MapColoringCSP(list('RGBY'),\n",
+ " \"\"\"WA: OR ID; OR: ID NV CA; CA: NV AZ; NV: ID UT AZ; ID: MT WY UT;\n",
+ " UT: WY CO AZ; MT: ND SD WY; WY: SD NE CO; CO: NE KA OK NM; NM: OK TX AZ;\n",
+ " ND: MN SD; SD: MN IA NE; NE: IA MO KA; KA: MO OK; OK: MO AR TX;\n",
+ " TX: AR LA; MN: WI IA; IA: WI IL MO; MO: IL KY TN AR; AR: MS TN LA;\n",
+ " LA: MS; WI: MI IL; IL: IN KY; IN: OH KY; MS: TN AL; AL: TN GA FL;\n",
+ " MI: OH IN; OH: PA WV KY; KY: WV VA TN; TN: VA NC GA; GA: NC SC FL;\n",
+ " PA: NY NJ DE MD WV; WV: MD VA; VA: MD DC NC; NC: SC; NY: VT MA CT NJ;\n",
+ " NJ: DE; DE: MD; MD: DC; VT: NH MA; MA: NH RI CT; CT: RI; ME: NH;\n",
+ " HI: ; AK: \"\"\")"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 54,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 1.58 ms, sys: 17 µs, total: 1.6 ms\n",
+ "Wall time: 1.6 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b with DOM J UP needs 1284 consistency-checks'"
+ ]
+ },
+ "execution_count": 54,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, checks = AC3b(usa_csp, arc_heuristic=dom_j_up)\n",
+ "f'AC3b with DOM J UP needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 55,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 2.15 ms, sys: 0 ns, total: 2.15 ms\n",
+ "Wall time: 2.15 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "{'NM': 'R',\n",
+ " 'TX': 'G',\n",
+ " 'OK': 'B',\n",
+ " 'AR': 'R',\n",
+ " 'MO': 'G',\n",
+ " 'KA': 'R',\n",
+ " 'LA': 'B',\n",
+ " 'NE': 'B',\n",
+ " 'TN': 'B',\n",
+ " 'MS': 'G',\n",
+ " 'IA': 'R',\n",
+ " 'SD': 'G',\n",
+ " 'IL': 'B',\n",
+ " 'CO': 'G',\n",
+ " 'MN': 'B',\n",
+ " 'KY': 'R',\n",
+ " 'AL': 'R',\n",
+ " 'GA': 'G',\n",
+ " 'FL': 'B',\n",
+ " 'VA': 'G',\n",
+ " 'WI': 'G',\n",
+ " 'IN': 'G',\n",
+ " 'NC': 'R',\n",
+ " 'WV': 'B',\n",
+ " 'OH': 'Y',\n",
+ " 'PA': 'R',\n",
+ " 'MD': 'Y',\n",
+ " 'SC': 'B',\n",
+ " 'MI': 'R',\n",
+ " 'DC': 'R',\n",
+ " 'DE': 'G',\n",
+ " 'WY': 'R',\n",
+ " 'ND': 'R',\n",
+ " 'NJ': 'B',\n",
+ " 'NY': 'G',\n",
+ " 'UT': 'B',\n",
+ " 'AZ': 'G',\n",
+ " 'ID': 'G',\n",
+ " 'MT': 'B',\n",
+ " 'NV': 'R',\n",
+ " 'CA': 'B',\n",
+ " 'OR': 'Y',\n",
+ " 'WA': 'R',\n",
+ " 'VT': 'R',\n",
+ " 'MA': 'B',\n",
+ " 'NH': 'G',\n",
+ " 'CT': 'R',\n",
+ " 'RI': 'G',\n",
+ " 'ME': 'R'}"
+ ]
+ },
+ "execution_count": 55,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time backtracking_search(usa_csp, select_unassigned_variable=mrv, inference=forward_checking)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### SAT"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 56,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "usa_sat = MapColoringSAT(list('RGBY'),\n",
+ " \"\"\"WA: OR ID; OR: ID NV CA; CA: NV AZ; NV: ID UT AZ; ID: MT WY UT;\n",
+ " UT: WY CO AZ; MT: ND SD WY; WY: SD NE CO; CO: NE KA OK NM; NM: OK TX AZ;\n",
+ " ND: MN SD; SD: MN IA NE; NE: IA MO KA; KA: MO OK; OK: MO AR TX;\n",
+ " TX: AR LA; MN: WI IA; IA: WI IL MO; MO: IL KY TN AR; AR: MS TN LA;\n",
+ " LA: MS; WI: MI IL; IL: IN KY; IN: OH KY; MS: TN AL; AL: TN GA FL;\n",
+ " MI: OH IN; OH: PA WV KY; KY: WV VA TN; TN: VA NC GA; GA: NC SC FL;\n",
+ " PA: NY NJ DE MD WV; WV: MD VA; VA: MD DC NC; NC: SC; NY: VT MA CT NJ;\n",
+ " NJ: DE; DE: MD; MD: DC; VT: NH MA; MA: NH RI CT; CT: RI; ME: NH;\n",
+ " HI: ; AK: \"\"\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "##### DPLL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 57,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 46.2 s, sys: 0 ns, total: 46.2 s\n",
+ "Wall time: 46.2 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(usa_sat, branching_heuristic=no_branching_heuristic)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 58,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 54.6 s, sys: 0 ns, total: 54.6 s\n",
+ "Wall time: 54.6 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(usa_sat, branching_heuristic=moms)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 59,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 44 s, sys: 0 ns, total: 44 s\n",
+ "Wall time: 44 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(usa_sat, branching_heuristic=momsf)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 60,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 43.8 s, sys: 0 ns, total: 43.8 s\n",
+ "Wall time: 43.8 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(usa_sat, branching_heuristic=posit)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 61,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 52.6 s, sys: 0 ns, total: 52.6 s\n",
+ "Wall time: 52.6 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(usa_sat, branching_heuristic=zm)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 62,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 57 s, sys: 0 ns, total: 57 s\n",
+ "Wall time: 57 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(usa_sat, branching_heuristic=dlis)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 63,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 43.8 s, sys: 0 ns, total: 43.8 s\n",
+ "Wall time: 43.8 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(usa_sat, branching_heuristic=dlcs)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 64,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 53.3 s, sys: 3.82 ms, total: 53.3 s\n",
+ "Wall time: 53.3 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(usa_sat, branching_heuristic=jw)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 65,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 44 s, sys: 3.99 ms, total: 44 s\n",
+ "Wall time: 44 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(usa_sat, branching_heuristic=jw2)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "##### CDCL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 66,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 559 ms, sys: 0 ns, total: 559 ms\n",
+ "Wall time: 558 ms\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = cdcl_satisfiable(usa_sat)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 67,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "{AL_B,\n",
+ " AR_B,\n",
+ " AZ_R,\n",
+ " CA_B,\n",
+ " CO_R,\n",
+ " CT_Y,\n",
+ " DC_G,\n",
+ " DE_Y,\n",
+ " FL_Y,\n",
+ " GA_R,\n",
+ " IA_B,\n",
+ " ID_Y,\n",
+ " IL_G,\n",
+ " IN_R,\n",
+ " KA_G,\n",
+ " KY_B,\n",
+ " LA_G,\n",
+ " MA_G,\n",
+ " MD_R,\n",
+ " ME_G,\n",
+ " MI_G,\n",
+ " MN_Y,\n",
+ " MO_R,\n",
+ " MS_Y,\n",
+ " MT_B,\n",
+ " NC_B,\n",
+ " ND_G,\n",
+ " NE_Y,\n",
+ " NH_Y,\n",
+ " NJ_G,\n",
+ " NM_G,\n",
+ " NV_G,\n",
+ " NY_R,\n",
+ " OH_Y,\n",
+ " OK_Y,\n",
+ " OR_R,\n",
+ " PA_B,\n",
+ " RI_B,\n",
+ " SC_Y,\n",
+ " SD_R,\n",
+ " TN_G,\n",
+ " TX_R,\n",
+ " UT_B,\n",
+ " VA_Y,\n",
+ " VT_B,\n",
+ " WA_B,\n",
+ " WI_R,\n",
+ " WV_G,\n",
+ " WY_G}"
+ ]
+ },
+ "execution_count": 67,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "{var for var, val in model.items() if val}"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Zebra Puzzle"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### CSP"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 76,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "zebra_csp = Zebra()"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 77,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "{'Milk': 3, 'Norwegian': 1}\n"
+ ]
+ }
+ ],
+ "source": [
+ "zebra_csp.display(zebra_csp.infer_assignment())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 78,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 2.04 ms, sys: 4 µs, total: 2.05 ms\n",
+ "Wall time: 2.05 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "'AC3b with DOM J UP needs 737 consistency-checks'"
+ ]
+ },
+ "execution_count": 78,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time _, checks = AC3b(zebra_csp, arc_heuristic=dom_j_up)\n",
+ "f'AC3b with DOM J UP needs {checks} consistency-checks'"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 71,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "{'Blue': 2, 'Milk': 3, 'Norwegian': 1}\n"
+ ]
+ }
+ ],
+ "source": [
+ "zebra_csp.display(zebra_csp.infer_assignment())"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 72,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 2.13 ms, sys: 0 ns, total: 2.13 ms\n",
+ "Wall time: 2.14 ms\n"
+ ]
+ },
+ {
+ "data": {
+ "text/plain": [
+ "{'Milk': 3,\n",
+ " 'Blue': 2,\n",
+ " 'Norwegian': 1,\n",
+ " 'Coffee': 5,\n",
+ " 'Green': 5,\n",
+ " 'Ivory': 4,\n",
+ " 'Red': 3,\n",
+ " 'Yellow': 1,\n",
+ " 'Kools': 1,\n",
+ " 'Englishman': 3,\n",
+ " 'Horse': 2,\n",
+ " 'Tea': 2,\n",
+ " 'Ukranian': 2,\n",
+ " 'Spaniard': 4,\n",
+ " 'Dog': 4,\n",
+ " 'Japanese': 5,\n",
+ " 'Parliaments': 5,\n",
+ " 'LuckyStrike': 4,\n",
+ " 'OJ': 4,\n",
+ " 'Water': 1,\n",
+ " 'Chesterfields': 2,\n",
+ " 'Winston': 3,\n",
+ " 'Snails': 3,\n",
+ " 'Fox': 1,\n",
+ " 'Zebra': 5}"
+ ]
+ },
+ "execution_count": 72,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "%time backtracking_search(zebra_csp, select_unassigned_variable=mrv, inference=forward_checking)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### SAT"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 5,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "zebra_sat = associate('&', map(to_cnf, map(expr, filter(lambda line: line[0] not in ('c', 'p'), open('aima-data/zebra.cnf').read().splitlines()))))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "##### DPLL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 13min 6s, sys: 2.44 ms, total: 13min 6s\n",
+ "Wall time: 13min 6s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(zebra_sat, branching_heuristic=no_branching_heuristic)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 16,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 15min 4s, sys: 22.4 ms, total: 15min 4s\n",
+ "Wall time: 15min 4s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(zebra_sat, branching_heuristic=moms)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 17,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 22min 28s, sys: 40 ms, total: 22min 28s\n",
+ "Wall time: 22min 28s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(zebra_sat, branching_heuristic=momsf)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 18,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 22min 25s, sys: 36 ms, total: 22min 25s\n",
+ "Wall time: 22min 25s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(zebra_sat, branching_heuristic=posit)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 14min 52s, sys: 32 ms, total: 14min 52s\n",
+ "Wall time: 14min 52s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(zebra_sat, branching_heuristic=zm)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 2min 31s, sys: 9.87 ms, total: 2min 31s\n",
+ "Wall time: 2min 32s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(zebra_sat, branching_heuristic=dlis)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 21,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 4min 27s, sys: 12 ms, total: 4min 27s\n",
+ "Wall time: 4min 27s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(zebra_sat, branching_heuristic=dlcs)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 6min 55s, sys: 39.2 ms, total: 6min 55s\n",
+ "Wall time: 6min 56s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(zebra_sat, branching_heuristic=jw)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 75,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 8min 57s, sys: 7.94 ms, total: 8min 57s\n",
+ "Wall time: 8min 57s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = dpll_satisfiable(zebra_sat, branching_heuristic=jw2)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "##### CDCL"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {
+ "pycharm": {}
+ },
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "CPU times: user 1.64 s, sys: 0 ns, total: 1.64 s\n",
+ "Wall time: 1.64 s\n"
+ ]
+ }
+ ],
+ "source": [
+ "%time model = cdcl_satisfiable(zebra_sat)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "{Englishman_house2,\n",
+ " Englishman_milk,\n",
+ " Englishman_oldGold,\n",
+ " Englishman_redHouse,\n",
+ " Englishman_snails,\n",
+ " Japanese_coffee,\n",
+ " Japanese_greenHouse,\n",
+ " Japanese_house4,\n",
+ " Japanese_parliament,\n",
+ " Japanese_zebra,\n",
+ " Norwegian_fox,\n",
+ " Norwegian_house0,\n",
+ " Norwegian_kool,\n",
+ " Norwegian_water,\n",
+ " Norwegian_yellowHouse,\n",
+ " Spaniard_dog,\n",
+ " Spaniard_house3,\n",
+ " Spaniard_ivoryHouse,\n",
+ " Spaniard_luckyStrike,\n",
+ " Spaniard_orangeJuice,\n",
+ " Ukrainian_blueHouse,\n",
+ " Ukrainian_chesterfield,\n",
+ " Ukrainian_horse,\n",
+ " Ukrainian_house1,\n",
+ " Ukrainian_tea}"
+ ]
+ },
+ "execution_count": 8,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "{var for var, val in model.items() if val and var.op.startswith(('Englishman', 'Japanese', 'Norwegian', 'Spaniard', 'Ukrainian'))}"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## References\n",
+ "\n",
+ "[[1]](#ref-1) Freeman, Jon William. 1995. _Improvements to propositional satisfiability search algorithms_.\n",
+ "\n",
+ "[[2]](#ref-2) Zabih, Ramin and McAllester, David A. 1988. _A Rearrangement Search Strategy for Determining Propositional Satisfiability_.\n",
+ "\n",
+ "[[3]](#ref-3) Jeroslow, Robert G and Wang, Jinchang. 1990. _Solving propositional satisfiability problems_.\n",
+ "\n",
+ "[[4]](#ref-4) Moskewicz, Matthew W and Madigan, Conor F and Zhao, Ying and Zhang, Lintao and Malik, Sharad. 2001. _Chaff: Engineering an efficient SAT solver_.\n",
+ "\n",
+ "[[5]](#ref-5) Haim, Shai and Heule, Marijn. 2014. _Towards ultra rapid restarts_.\n",
+ "\n",
+ "[[6]](#ref-6) Huang, Jinbo and others. 2007. _The Effect of Restarts on the Efficiency of Clause Learning_.\n",
+ "\n",
+ "[[7]](#ref-7) Audemard, Gilles and Simon, Laurent. 2012. _Refining restarts strategies for SAT and UNSAT_.\n",
+ "\n",
+ "[[8]](#ref-8) Audemard, Gilles and Simon, Laurent. 2009. _Predicting learnt clauses quality in modern SAT solvers_."
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.7.3"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}
diff --git a/knowledge.py b/knowledge.py
index eaeacf7d9..a33eac81a 100644
--- a/knowledge.py
+++ b/knowledge.py
@@ -14,7 +14,8 @@
def current_best_learning(examples, h, examples_so_far=None):
- """ [Figure 19.2]
+ """
+ [Figure 19.2]
The hypothesis is a list of dictionaries, with each dictionary representing
a disjunction."""
if examples_so_far is None:
@@ -124,7 +125,8 @@ def add_or(examples_so_far, h):
def version_space_learning(examples):
- """ [Figure 19.3]
+ """
+ [Figure 19.3]
The version space is a list of hypotheses, which in turn are a list
of dictionaries/disjunctions."""
V = all_hypotheses(examples)
@@ -241,7 +243,7 @@ def consistent_det(A, E):
# ______________________________________________________________________________
-class FOIL_container(FolKB):
+class FOILContainer(FolKB):
"""Hold the kb and other necessary elements required by FOIL."""
def __init__(self, clauses=None):
@@ -255,7 +257,7 @@ def tell(self, sentence):
self.const_syms.update(constant_symbols(sentence))
self.pred_syms.update(predicate_symbols(sentence))
else:
- raise Exception("Not a definite clause: {}".format(sentence))
+ raise Exception('Not a definite clause: {}'.format(sentence))
def foil(self, examples, target):
"""Learn a list of first-order horn clauses
@@ -280,7 +282,6 @@ def new_clause(self, examples, target):
The horn clause is specified as [consequent, list of antecedents]
Return value is the tuple (horn_clause, extended_positive_examples)."""
clause = [target, []]
- # [positive_examples, negative_examples]
extended_examples = examples
while extended_examples[1]:
l = self.choose_literal(self.new_literals(clause), extended_examples)
@@ -288,7 +289,7 @@ def new_clause(self, examples, target):
extended_examples = [sum([list(self.extend_example(example, l)) for example in
extended_examples[i]], []) for i in range(2)]
- return (clause, extended_examples[0])
+ return clause, extended_examples[0]
def extend_example(self, example, literal):
"""Generate extended examples which satisfy the literal."""
@@ -344,9 +345,8 @@ def gain(self, l, examples):
represents = lambda d: all(d[x] == example[x] for x in example)
if any(represents(l_) for l_ in post_pos):
T += 1
- value = T * (
- log(len(post_pos) / (len(post_pos) + len(post_neg)) + 1e-12, 2) - log(pre_pos / (pre_pos + pre_neg),
- 2))
+ value = T * (log(len(post_pos) / (len(post_pos) + len(post_neg)) + 1e-12, 2) -
+ log(pre_pos / (pre_pos + pre_neg), 2))
return value
def update_examples(self, target, examples, extended_examples):
@@ -411,12 +411,12 @@ def guess_value(e, h):
def is_consistent(e, h):
- return e["GOAL"] == guess_value(e, h)
+ return e['GOAL'] == guess_value(e, h)
def false_positive(e, h):
- return guess_value(e, h) and not e["GOAL"]
+ return guess_value(e, h) and not e['GOAL']
def false_negative(e, h):
- return e["GOAL"] and not guess_value(e, h)
+ return e['GOAL'] and not guess_value(e, h)
diff --git a/learning.py b/learning.py
index 31aabe30f..2d4bd4d4b 100644
--- a/learning.py
+++ b/learning.py
@@ -8,7 +8,7 @@
from statistics import mean, stdev
from probabilistic_learning import NaiveBayesLearner
-from utils import (remove_all, unique, mode, argmax, argmax_random_tie, isclose, dotproduct, vector_add,
+from utils import (remove_all, unique, mode, argmax, argmax_random_tie, isclose, dot_product, vector_add,
scalar_vector_product, weighted_sample_with_replacement, num_or_str, normalize, clip, sigmoid,
print_table, open_data, sigmoid_derivative, probability, relu, relu_derivative, tanh,
tanh_derivative, leaky_relu_derivative, elu, elu_derivative, mean_boolean_error, random_weights)
@@ -536,17 +536,17 @@ def LinearLearner(dataset, learning_rate=0.01, epochs=100):
# pass over all examples
for example in examples:
x = [1] + example
- y = dotproduct(w, x)
+ y = dot_product(w, x)
t = example[idx_t]
err.append(t - y)
# update weights
for i in range(len(w)):
- w[i] = w[i] + learning_rate * (dotproduct(err, X_col[i]) / num_examples)
+ w[i] = w[i] + learning_rate * (dot_product(err, X_col[i]) / num_examples)
def predict(example):
x = [1] + example
- return dotproduct(w, x)
+ return dot_product(w, x)
return predict
@@ -578,7 +578,7 @@ def LogisticLinearLeaner(dataset, learning_rate=0.01, epochs=100):
# pass over all examples
for example in examples:
x = [1] + example
- y = sigmoid(dotproduct(w, x))
+ y = sigmoid(dot_product(w, x))
h.append(sigmoid_derivative(y))
t = example[idx_t]
err.append(t - y)
@@ -586,11 +586,11 @@ def LogisticLinearLeaner(dataset, learning_rate=0.01, epochs=100):
# update weights
for i in range(len(w)):
buffer = [x * y for x, y in zip(err, h)]
- w[i] = w[i] + learning_rate * (dotproduct(buffer, X_col[i]) / num_examples)
+ w[i] = w[i] + learning_rate * (dot_product(buffer, X_col[i]) / num_examples)
def predict(example):
x = [1] + example
- return sigmoid(dotproduct(w, x))
+ return sigmoid(dot_product(w, x))
return predict
@@ -624,7 +624,7 @@ def predict(example):
for layer in learned_net[1:]:
for node in layer:
inc = [n.value for n in node.inputs]
- in_val = dotproduct(inc, node.weights)
+ in_val = dot_product(inc, node.weights)
node.value = node.activation(in_val)
# hypothesis
@@ -672,7 +672,7 @@ def BackPropagationLearner(dataset, net, learning_rate, epochs, activation=sigmo
for layer in net[1:]:
for node in layer:
inc = [n.value for n in node.inputs]
- in_val = dotproduct(inc, node.weights)
+ in_val = dot_product(inc, node.weights)
node.value = node.activation(in_val)
# initialize delta
@@ -706,19 +706,19 @@ def BackPropagationLearner(dataset, net, learning_rate, epochs, activation=sigmo
w = [[node.weights[k] for node in nx_layer] for k in range(h_units)]
if activation == sigmoid:
- delta[i] = [sigmoid_derivative(layer[j].value) * dotproduct(w[j], delta[i + 1])
+ delta[i] = [sigmoid_derivative(layer[j].value) * dot_product(w[j], delta[i + 1])
for j in range(h_units)]
elif activation == relu:
- delta[i] = [relu_derivative(layer[j].value) * dotproduct(w[j], delta[i + 1])
+ delta[i] = [relu_derivative(layer[j].value) * dot_product(w[j], delta[i + 1])
for j in range(h_units)]
elif activation == tanh:
- delta[i] = [tanh_derivative(layer[j].value) * dotproduct(w[j], delta[i + 1])
+ delta[i] = [tanh_derivative(layer[j].value) * dot_product(w[j], delta[i + 1])
for j in range(h_units)]
elif activation == elu:
- delta[i] = [elu_derivative(layer[j].value) * dotproduct(w[j], delta[i + 1])
+ delta[i] = [elu_derivative(layer[j].value) * dot_product(w[j], delta[i + 1])
for j in range(h_units)]
else:
- delta[i] = [leaky_relu_derivative(layer[j].value) * dotproduct(w[j], delta[i + 1])
+ delta[i] = [leaky_relu_derivative(layer[j].value) * dot_product(w[j], delta[i + 1])
for j in range(h_units)]
# update weights
@@ -746,7 +746,7 @@ def predict(example):
# forward pass
for node in o_nodes:
- in_val = dotproduct(example, node.weights)
+ in_val = dot_product(example, node.weights)
node.value = node.activation(in_val)
# hypothesis
diff --git a/learning4e.py b/learning4e.py
index 5cf63dda4..e4a566667 100644
--- a/learning4e.py
+++ b/learning4e.py
@@ -1,4 +1,4 @@
-"""Learning from examples. (Chapters 18)"""
+"""Learning from examples (Chapters 18)"""
import copy
import heapq
@@ -9,9 +9,9 @@
from probabilistic_learning import NaiveBayesLearner
from utils import sigmoid, sigmoid_derivative
-from utils4e import (remove_all, unique, mode, argmax_random_tie, isclose, dotproduct, weighted_sample_with_replacement,
- num_or_str, normalize, clip, print_table, open_data, probability, random_weights,
- mean_boolean_error)
+from utils4e import (remove_all, unique, mode, argmax_random_tie, isclose, dot_product,
+ weighted_sample_with_replacement, num_or_str, normalize, clip, print_table, open_data, probability,
+ random_weights, mean_boolean_error)
class DataSet:
@@ -531,17 +531,17 @@ def LinearLearner(dataset, learning_rate=0.01, epochs=100):
# pass over all examples
for example in examples:
x = [1] + example
- y = dotproduct(w, x)
+ y = dot_product(w, x)
t = example[idx_t]
err.append(t - y)
# update weights
for i in range(len(w)):
- w[i] = w[i] + learning_rate * (dotproduct(err, X_col[i]) / num_examples)
+ w[i] = w[i] + learning_rate * (dot_product(err, X_col[i]) / num_examples)
def predict(example):
x = [1] + example
- return dotproduct(w, x)
+ return dot_product(w, x)
return predict
@@ -573,7 +573,7 @@ def LogisticLinearLeaner(dataset, learning_rate=0.01, epochs=100):
# pass over all examples
for example in examples:
x = [1] + example
- y = sigmoid(dotproduct(w, x))
+ y = sigmoid(dot_product(w, x))
h.append(sigmoid_derivative(y))
t = example[idx_t]
err.append(t - y)
@@ -581,11 +581,11 @@ def LogisticLinearLeaner(dataset, learning_rate=0.01, epochs=100):
# update weights
for i in range(len(w)):
buffer = [x * y for x, y in zip(err, h)]
- w[i] = w[i] + learning_rate * (dotproduct(buffer, X_col[i]) / num_examples)
+ w[i] = w[i] + learning_rate * (dot_product(buffer, X_col[i]) / num_examples)
def predict(example):
x = [1] + example
- return sigmoid(dotproduct(w, x))
+ return sigmoid(dot_product(w, x))
return predict
diff --git a/logic.py b/logic.py
index 7f4d259dd..ae987edb4 100644
--- a/logic.py
+++ b/logic.py
@@ -1,4 +1,5 @@
-"""Representations and Inference for Logic (Chapters 7-9, 12)
+"""
+Representations and Inference for Logic (Chapters 7-9, 12)
Covers both Propositional and First-Order Logic. First we have four
important data types:
@@ -30,6 +31,7 @@
unify Do unification of two FOL sentences
diff, simp Symbolic differentiation and simplification
"""
+
import heapq
import itertools
import random
@@ -111,25 +113,28 @@ def retract(self, sentence):
# ______________________________________________________________________________
-def KB_AgentProgram(KB):
- """A generic logical knowledge-based agent program. [Figure 7.1]"""
+def KBAgentProgram(kb):
+ """
+ [Figure 7.1]
+ A generic logical knowledge-based agent program.
+ """
steps = itertools.count()
def program(percept):
t = next(steps)
- KB.tell(make_percept_sentence(percept, t))
- action = KB.ask(make_action_query(t))
- KB.tell(make_action_sentence(action, t))
+ kb.tell(make_percept_sentence(percept, t))
+ action = kb.ask(make_action_query(t))
+ kb.tell(make_action_sentence(action, t))
return action
def make_percept_sentence(percept, t):
- return Expr("Percept")(percept, t)
+ return Expr('Percept')(percept, t)
def make_action_query(t):
- return expr("ShouldDo(action, {})".format(t))
+ return expr('ShouldDo(action, {})'.format(t))
def make_action_sentence(action, t):
- return Expr("Did")(action[expr('action')], t)
+ return Expr('Did')(action[expr('action')], t)
return program
@@ -177,8 +182,7 @@ def is_definite_clause(s):
return True
elif s.op == '==>':
antecedent, consequent = s.args
- return (is_symbol(consequent.op) and
- all(is_symbol(arg.op) for arg in conjuncts(antecedent)))
+ return is_symbol(consequent.op) and all(is_symbol(arg.op) for arg in conjuncts(antecedent))
else:
return False
@@ -201,9 +205,11 @@ def parse_definite_clause(s):
def tt_entails(kb, alpha):
- """Does kb entail the sentence alpha? Use truth tables. For propositional
- kb's and sentences. [Figure 7.10]. Note that the 'kb' should be an
- Expr which is a conjunction of clauses.
+ """
+ [Figure 7.10]
+ Does kb entail the sentence alpha? Use truth tables. For propositional
+ kb's and sentences. Note that the 'kb' should be an Expr which is a
+ conjunction of clauses.
>>> tt_entails(expr('P & Q'), expr('Q'))
True
"""
@@ -319,7 +325,7 @@ def pl_true(exp, model={}):
elif op == '^': # xor or 'not equivalent'
return pt != qt
else:
- raise ValueError("illegal operator in logic expression" + str(exp))
+ raise ValueError('Illegal operator in logic expression' + str(exp))
# ______________________________________________________________________________
@@ -328,8 +334,10 @@ def pl_true(exp, model={}):
def to_cnf(s):
- """Convert a propositional logical sentence to conjunctive normal form.
- That is, to the form ((A | ~B | ...) & (B | C | ...) & ...) [p. 253]
+ """
+ [Page 253]
+ Convert a propositional logical sentence to conjunctive normal form.
+ That is, to the form ((A | ~B | ...) & (B | C | ...) & ...)
>>> to_cnf('~(B | C)')
(~B & ~C)
"""
@@ -477,12 +485,14 @@ def disjuncts(s):
# ______________________________________________________________________________
-def pl_resolution(KB, alpha):
- """Propositional-logic resolution: say if alpha follows from KB. [Figure 7.12]
+def pl_resolution(kb, alpha):
+ """
+ [Figure 7.12]
+ Propositional-logic resolution: say if alpha follows from KB.
>>> pl_resolution(horn_clauses_KB, A)
True
"""
- clauses = KB.clauses + conjuncts(to_cnf(~alpha))
+ clauses = kb.clauses + conjuncts(to_cnf(~alpha))
new = set()
while True:
n = len(clauses)
@@ -532,52 +542,62 @@ def retract(self, sentence):
def clauses_with_premise(self, p):
"""Return a list of the clauses in KB that have p in their premise.
This could be cached away for O(1) speed, but we'll recompute it."""
- return [c for c in self.clauses
- if c.op == '==>' and p in conjuncts(c.args[0])]
+ return [c for c in self.clauses if c.op == '==>' and p in conjuncts(c.args[0])]
-def pl_fc_entails(KB, q):
- """Use forward chaining to see if a PropDefiniteKB entails symbol q.
+def pl_fc_entails(kb, q):
+ """
[Figure 7.15]
+ Use forward chaining to see if a PropDefiniteKB entails symbol q.
>>> pl_fc_entails(horn_clauses_KB, expr('Q'))
True
"""
- count = {c: len(conjuncts(c.args[0]))
- for c in KB.clauses
- if c.op == '==>'}
+ count = {c: len(conjuncts(c.args[0])) for c in kb.clauses if c.op == '==>'}
inferred = defaultdict(bool)
- agenda = [s for s in KB.clauses if is_prop_symbol(s.op)]
+ agenda = [s for s in kb.clauses if is_prop_symbol(s.op)]
while agenda:
p = agenda.pop()
if p == q:
return True
if not inferred[p]:
inferred[p] = True
- for c in KB.clauses_with_premise(p):
+ for c in kb.clauses_with_premise(p):
count[c] -= 1
if count[c] == 0:
agenda.append(c.args[1])
return False
-""" [Figure 7.13]
+"""
+[Figure 7.13]
Simple inference in a wumpus world example
"""
-wumpus_world_inference = expr("(B11 <=> (P12 | P21)) & ~B11")
+wumpus_world_inference = expr('(B11 <=> (P12 | P21)) & ~B11')
-""" [Figure 7.16]
+"""
+[Figure 7.16]
Propositional Logic Forward Chaining example
"""
horn_clauses_KB = PropDefiniteKB()
-for s in "P==>Q; (L&M)==>P; (B&L)==>M; (A&P)==>L; (A&B)==>L; A;B".split(';'):
- horn_clauses_KB.tell(expr(s))
+for clause in ['P ==> Q',
+ '(L & M) ==> P',
+ '(B & L) ==> M',
+ '(A & P) ==> L',
+ '(A & B) ==> L',
+ 'A', 'B']:
+ horn_clauses_KB.tell(expr(clause))
"""
Definite clauses KB example
"""
definite_clauses_KB = PropDefiniteKB()
-for clause in ['(B & F)==>E', '(A & E & F)==>G', '(B & C)==>F', '(A & B)==>D', '(E & F)==>H', '(H & I)==>J', 'A', 'B',
- 'C']:
+for clause in ['(B & F) ==> E',
+ '(A & E & F) ==> G',
+ '(B & C) ==> F',
+ '(A & B) ==> D',
+ '(E & F) ==> H',
+ '(H & I) ==>J',
+ 'A', 'B', 'C']:
definite_clauses_KB.tell(expr(clause))
@@ -1378,22 +1398,14 @@ def add_temporal_sentences(self, time):
for j in range(1, self.dimrow + 1):
self.tell(implies(location(i, j, time), equiv(percept_breeze(time), breeze(i, j))))
self.tell(implies(location(i, j, time), equiv(percept_stench(time), stench(i, j))))
-
s = list()
-
- s.append(
- equiv(
- location(i, j, time), location(i, j, time) & ~move_forward(time) | percept_bump(time)))
-
+ s.append(equiv(location(i, j, time), location(i, j, time) & ~move_forward(time) | percept_bump(time)))
if i != 1:
s.append(location(i - 1, j, t) & facing_east(t) & move_forward(t))
-
if i != self.dimrow:
s.append(location(i + 1, j, t) & facing_west(t) & move_forward(t))
-
if j != 1:
s.append(location(i, j - 1, t) & facing_north(t) & move_forward(t))
-
if j != self.dimrow:
s.append(location(i, j + 1, t) & facing_south(t) & move_forward(t))
@@ -1401,9 +1413,7 @@ def add_temporal_sentences(self, time):
self.tell(new_disjunction(s))
# add sentence about safety of location i,j
- self.tell(
- equiv(ok_to_move(i, j, time), ~pit(i, j) & ~wumpus(i, j) & wumpus_alive(time))
- )
+ self.tell(equiv(ok_to_move(i, j, time), ~pit(i, j) & ~wumpus(i, j) & wumpus_alive(time)))
# Rules about current orientation
@@ -1477,7 +1487,10 @@ def __eq__(self, other):
class HybridWumpusAgent(Agent):
- """An agent for the wumpus world that does logical inference. [Figure 7.20]"""
+ """
+ [Figure 7.20]
+ An agent for the wumpus world that does logical inference.
+ """
def __init__(self, dimentions):
self.dimrow = dimentions
@@ -1607,8 +1620,9 @@ def plan_shot(self, current, goals, allowed):
def SAT_plan(init, transition, goal, t_max, SAT_solver=cdcl_satisfiable):
- """Converts a planning problem to Satisfaction problem by translating it to a cnf sentence.
+ """
[Figure 7.22]
+ Converts a planning problem to Satisfaction problem by translating it to a cnf sentence.
>>> transition = {'A': {'Left': 'A', 'Right': 'B'}, 'B': {'Left': 'A', 'Right': 'C'}, 'C': {'Left': 'B', 'Right': 'C'}}
>>> SAT_plan('A', transition, 'C', 1) is None
True
@@ -1623,7 +1637,7 @@ def translate_to_SAT(init, transition, goal, time):
state_counter = itertools.count()
for s in states:
for t in range(time + 1):
- state_sym[s, t] = Expr("S{}".format(next(state_counter)))
+ state_sym[s, t] = Expr('S_{}'.format(next(state_counter)))
# Add initial state axiom
clauses.append(state_sym[init, 0])
@@ -1640,7 +1654,7 @@ def translate_to_SAT(init, transition, goal, time):
s_ = transition[s][action]
for t in range(time):
# Action 'action' taken from state 's' at time 't' to reach 's_'
- action_sym[s, action, t] = Expr("T{}".format(next(transition_counter)))
+ action_sym[s, action, t] = Expr('T_{}'.format(next(transition_counter)))
# Change the state from s to s_
clauses.append(action_sym[s, action, t] | '==>' | state_sym[s, t])
@@ -1695,9 +1709,11 @@ def extract_solution(model):
def unify(x, y, s={}):
- """Unify expressions x,y with substitution s; return a substitution that
+ """
+ [Figure 9.1]
+ Unify expressions x,y with substitution s; return a substitution that
would make x,y equal, or None if x,y can not unify. x and y can be
- variables (e.g. Expr('x')), constants, lists, or Exprs. [Figure 9.1]
+ variables (e.g. Expr('x')), constants, lists, or Exprs.
>>> unify(x, 3, {})
{x: 3}
"""
@@ -1791,6 +1807,80 @@ def cascade_substitution(s):
s[x] = subst(s, s.get(x))
+def unify_mm(x, y, s={}):
+ """Unify expressions x,y with substitution s using an efficient rule-based
+ unification algorithm by Martelli & Montanari; return a substitution that
+ would make x,y equal, or None if x,y can not unify. x and y can be
+ variables (e.g. Expr('x')), constants, lists, or Exprs.
+ >>> unify_mm(x, 3, {})
+ {x: 3}
+ """
+
+ set_eq = extend(s, x, y)
+ s = set_eq.copy()
+ while True:
+ trans = 0
+ for x, y in set_eq.items():
+ if x == y:
+ # if x = y this mapping is deleted (rule b)
+ del s[x]
+ elif not is_variable(x) and is_variable(y):
+ # if x is not a variable and y is a variable, rewrite it as y = x in s (rule a)
+ if s.get(y, None) is None:
+ s[y] = x
+ del s[x]
+ else:
+ # if a mapping already exist for variable y then apply
+ # variable elimination (there is a chance to apply rule d)
+ s[x] = vars_elimination(y, s)
+ elif not is_variable(x) and not is_variable(y):
+ # in which case x and y are not variables, if the two root function symbols
+ # are different, stop with failure, else apply term reduction (rule c)
+ if x.op is y.op and len(x.args) == len(y.args):
+ term_reduction(x, y, s)
+ del s[x]
+ else:
+ return None
+ elif isinstance(y, Expr):
+ # in which case x is a variable and y is a function or a variable (e.g. F(z) or y),
+ # if y is a function, we must check if x occurs in y, then stop with failure, else
+ # try to apply variable elimination to y (rule d)
+ if occur_check(x, y, s):
+ return None
+ s[x] = vars_elimination(y, s)
+ if y == s.get(x):
+ trans += 1
+ else:
+ trans += 1
+ if trans == len(set_eq):
+ # if no transformation has been applied, stop with success
+ return s
+ set_eq = s.copy()
+
+
+def term_reduction(x, y, s):
+ """Apply term reduction to x and y if both are functions and the two root function
+ symbols are equals (e.g. F(x1, x2, ..., xn) and F(x1', x2', ..., xn')) by returning
+ a new mapping obtained by replacing x: y with {x1: x1', x2: x2', ..., xn: xn'}
+ """
+ for i in range(len(x.args)):
+ if x.args[i] in s:
+ s[s.get(x.args[i])] = y.args[i]
+ else:
+ s[x.args[i]] = y.args[i]
+
+
+def vars_elimination(x, s):
+ """Apply variable elimination to x: if x is a variable and occurs in s, return
+ the term mapped by x, else if x is a function recursively applies variable
+ elimination to each term of the function."""
+ if not isinstance(x, Expr):
+ return x
+ if is_variable(x):
+ return s.get(x, x)
+ return Expr(x.op, *[vars_elimination(arg, s) for arg in x.args])
+
+
def standardize_variables(sentence, dic=None):
"""Replace all the variables in sentence with new variables."""
if dic is None:
@@ -1814,6 +1904,19 @@ def standardize_variables(sentence, dic=None):
# ______________________________________________________________________________
+def parse_clauses_from_dimacs(dimacs_cnf):
+ """Converts a string into CNF clauses according to the DIMACS format used in SAT competitions"""
+ return map(lambda c: associate('|', c),
+ map(lambda c: [expr('~X' + str(abs(l))) if l < 0 else expr('X' + str(l)) for l in c],
+ map(lambda line: map(int, line.split()),
+ filter(None, ' '.join(
+ filter(lambda line: line[0] not in ('c', 'p'),
+ filter(None, dimacs_cnf.strip().replace('\t', ' ').split('\n')))).split(' 0')))))
+
+
+# ______________________________________________________________________________
+
+
class FolKB(KB):
"""A knowledge base consisting of first-order definite clauses.
>>> kb0 = FolKB([expr('Farmer(Mac)'), expr('Rabbit(Pete)'),
@@ -1836,7 +1939,7 @@ def tell(self, sentence):
if is_definite_clause(sentence):
self.clauses.append(sentence)
else:
- raise Exception("Not a definite clause: {}".format(sentence))
+ raise Exception('Not a definite clause: {}'.format(sentence))
def ask_generator(self, query):
return fol_bc_ask(self, query)
@@ -1848,10 +1951,13 @@ def fetch_rules_for_goal(self, goal):
return self.clauses
-def fol_fc_ask(KB, alpha):
- """A simple forward-chaining algorithm. [Figure 9.3]"""
+def fol_fc_ask(kb, alpha):
+ """
+ [Figure 9.3]
+ A simple forward-chaining algorithm.
+ """
# TODO: Improve efficiency
- kb_consts = list({c for clause in KB.clauses for c in constant_symbols(clause)})
+ kb_consts = list({c for clause in kb.clauses for c in constant_symbols(clause)})
def enum_subst(p):
query_vars = list({v for clause in p for v in variables(clause)})
@@ -1860,19 +1966,19 @@ def enum_subst(p):
yield theta
# check if we can answer without new inferences
- for q in KB.clauses:
+ for q in kb.clauses:
phi = unify(q, alpha)
if phi is not None:
yield phi
while True:
new = []
- for rule in KB.clauses:
+ for rule in kb.clauses:
p, q = parse_definite_clause(rule)
for theta in enum_subst(p):
- if set(subst(theta, p)).issubset(set(KB.clauses)):
+ if set(subst(theta, p)).issubset(set(kb.clauses)):
q_ = subst(theta, q)
- if all([unify(x, q_) is None for x in KB.clauses + new]):
+ if all([unify(x, q_) is None for x in kb.clauses + new]):
new.append(q_)
phi = unify(q_, alpha)
if phi is not None:
@@ -1880,32 +1986,35 @@ def enum_subst(p):
if not new:
break
for clause in new:
- KB.tell(clause)
+ kb.tell(clause)
return None
-def fol_bc_ask(KB, query):
- """A simple backward-chaining algorithm for first-order logic. [Figure 9.6]
- KB should be an instance of FolKB, and query an atomic sentence."""
- return fol_bc_or(KB, query, {})
+def fol_bc_ask(kb, query):
+ """
+ [Figure 9.6]
+ A simple backward-chaining algorithm for first-order logic.
+ KB should be an instance of FolKB, and query an atomic sentence.
+ """
+ return fol_bc_or(kb, query, {})
-def fol_bc_or(KB, goal, theta):
- for rule in KB.fetch_rules_for_goal(goal):
+def fol_bc_or(kb, goal, theta):
+ for rule in kb.fetch_rules_for_goal(goal):
lhs, rhs = parse_definite_clause(standardize_variables(rule))
- for theta1 in fol_bc_and(KB, lhs, unify(rhs, goal, theta)):
+ for theta1 in fol_bc_and(kb, lhs, unify(rhs, goal, theta)):
yield theta1
-def fol_bc_and(KB, goals, theta):
+def fol_bc_and(kb, goals, theta):
if theta is None:
pass
elif not goals:
yield theta
else:
first, rest = goals[0], goals[1:]
- for theta1 in fol_bc_or(KB, subst(theta, first), theta):
- for theta2 in fol_bc_and(KB, rest, theta1):
+ for theta1 in fol_bc_or(kb, subst(theta, first), theta):
+ for theta2 in fol_bc_and(kb, rest, theta1):
yield theta2
@@ -1920,31 +2029,27 @@ def fol_bc_and(KB, goals, theta):
wumpus_kb.tell(~B11)
wumpus_kb.tell(B21)
-test_kb = FolKB(
- map(expr, ['Farmer(Mac)',
- 'Rabbit(Pete)',
- 'Mother(MrsMac, Mac)',
- 'Mother(MrsRabbit, Pete)',
- '(Rabbit(r) & Farmer(f)) ==> Hates(f, r)',
- '(Mother(m, c)) ==> Loves(m, c)',
- '(Mother(m, r) & Rabbit(r)) ==> Rabbit(m)',
- '(Farmer(f)) ==> Human(f)',
- # Note that this order of conjuncts
- # would result in infinite recursion:
- # '(Human(h) & Mother(m, h)) ==> Human(m)'
- '(Mother(m, h) & Human(h)) ==> Human(m)'
- ]))
-
-crime_kb = FolKB(
- map(expr, ['(American(x) & Weapon(y) & Sells(x, y, z) & Hostile(z)) ==> Criminal(x)',
- 'Owns(Nono, M1)',
- 'Missile(M1)',
- '(Missile(x) & Owns(Nono, x)) ==> Sells(West, x, Nono)',
- 'Missile(x) ==> Weapon(x)',
- 'Enemy(x, America) ==> Hostile(x)',
- 'American(West)',
- 'Enemy(Nono, America)'
- ]))
+test_kb = FolKB(map(expr, ['Farmer(Mac)',
+ 'Rabbit(Pete)',
+ 'Mother(MrsMac, Mac)',
+ 'Mother(MrsRabbit, Pete)',
+ '(Rabbit(r) & Farmer(f)) ==> Hates(f, r)',
+ '(Mother(m, c)) ==> Loves(m, c)',
+ '(Mother(m, r) & Rabbit(r)) ==> Rabbit(m)',
+ '(Farmer(f)) ==> Human(f)',
+ # Note that this order of conjuncts
+ # would result in infinite recursion:
+ # '(Human(h) & Mother(m, h)) ==> Human(m)'
+ '(Mother(m, h) & Human(h)) ==> Human(m)']))
+
+crime_kb = FolKB(map(expr, ['(American(x) & Weapon(y) & Sells(x, y, z) & Hostile(z)) ==> Criminal(x)',
+ 'Owns(Nono, M1)',
+ 'Missile(M1)',
+ '(Missile(x) & Owns(Nono, x)) ==> Sells(West, x, Nono)',
+ 'Missile(x) ==> Weapon(x)',
+ 'Enemy(x, America) ==> Hostile(x)',
+ 'American(West)',
+ 'Enemy(Nono, America)']))
# ______________________________________________________________________________
@@ -1984,7 +2089,7 @@ def diff(y, x):
elif op == 'log':
return diff(u, x) / u
else:
- raise ValueError("Unknown op: {} in diff({}, {})".format(op, y, x))
+ raise ValueError('Unknown op: {} in diff({}, {})'.format(op, y, x))
def simp(x):
@@ -2045,7 +2150,7 @@ def simp(x):
if u == 1:
return 0
else:
- raise ValueError("Unknown op: " + op)
+ raise ValueError('Unknown op: ' + op)
# If we fall through to here, we can not simplify further
return Expr(op, *args)
diff --git a/mdp4e.py b/mdp4e.py
index 5fadf2f67..bef1a7940 100644
--- a/mdp4e.py
+++ b/mdp4e.py
@@ -1,10 +1,12 @@
-"""Markov Decision Processes (Chapter 16)
+"""
+Markov Decision Processes (Chapter 16)
First we define an MDP, and the special case of a GridMDP, in which
states are laid out in a 2-dimensional grid. We also represent a policy
as a dictionary of {state: action} pairs, and a Utility function as a
dictionary of {state: number} pairs. We then define the value_iteration
-and policy_iteration algorithms."""
+and policy_iteration algorithms.
+"""
from utils4e import argmax, vector_add, orientations, turn_right, turn_left
from planning import *
diff --git a/perception4e.py b/perception4e.py
index 08238dfb7..887d014b2 100644
--- a/perception4e.py
+++ b/perception4e.py
@@ -3,7 +3,7 @@
import numpy as np
import scipy.signal
import matplotlib.pyplot as plt
-from utils4e import gaussian_kernel_2d
+from utils4e import gaussian_kernel_2d, inf
import keras
from keras.datasets import mnist
from keras.models import Sequential
@@ -86,8 +86,8 @@ def sum_squared_difference(pic1, pic2):
pic1 = np.asarray(pic1)
pic2 = np.asarray(pic2)
assert pic1.shape == pic2.shape
- min_ssd = float('inf')
- min_dxy = (float('inf'), float('inf'))
+ min_ssd = inf
+ min_dxy = (inf, inf)
# consider picture shift from -30 to 30
for Dx in range(-30, 31):
@@ -241,7 +241,7 @@ def min_cut(self, source, sink):
max_flow = 0
while self.bfs(source, sink, parent):
- path_flow = float('inf')
+ path_flow = inf
# find the minimum flow of s-t path
for s, t in parent:
path_flow = min(path_flow, self.flow[s][t])
diff --git a/planning.py b/planning.py
index b88b4f408..3835e05df 100644
--- a/planning.py
+++ b/planning.py
@@ -1,4 +1,5 @@
-"""Planning (Chapters 10-11)
+"""
+Planning (Chapters 10-11)
"""
import copy
@@ -10,7 +11,7 @@
from csp import sat_up, NaryCSP, Constraint, ac_search_solver, is_
from logic import FolKB, conjuncts, unify, associate, SAT_plan, cdcl_satisfiable
from search import Node
-from utils import Expr, expr, first
+from utils import Expr, expr, first, inf
class PlanningProblem:
@@ -316,7 +317,8 @@ def air_cargo():
def spare_tire():
- """[Figure 10.2] SPARE-TIRE-PROBLEM
+ """
+ [Figure 10.2] SPARE-TIRE-PROBLEM
A problem involving changing the flat tire of a car
with a spare tire from the trunk.
@@ -560,7 +562,8 @@ def double_tennis_problem():
class ForwardPlan(search.Problem):
"""
- Forward state-space search [Section 10.2.1]
+ [Section 10.2.1]
+ Forward state-space search
"""
def __init__(self, planning_problem):
@@ -580,7 +583,7 @@ def goal_test(self, state):
def h(self, state):
"""
Computes ignore delete lists heuristic by creating a relaxed version of the original problem (we can do that
- by removing the delete lists from all actions, ie. removing all negative literals from effects) that will be
+ by removing the delete lists from all actions, i.e. removing all negative literals from effects) that will be
easier to solve through GraphPlan and where the length of the solution will serve as a good heuristic.
"""
relaxed_planning_problem = PlanningProblem(initial=state.state,
@@ -590,12 +593,13 @@ def h(self, state):
try:
return len(linearize(GraphPlan(relaxed_planning_problem).execute()))
except:
- return float('inf')
+ return inf
class BackwardPlan(search.Problem):
"""
- Backward relevant-states search [Section 10.2.2]
+ [Section 10.2.2]
+ Backward relevant-states search
"""
def __init__(self, planning_problem):
@@ -605,7 +609,7 @@ def __init__(self, planning_problem):
def actions(self, subgoal):
"""
- Returns True if the action is relevant to the subgoal, ie.:
+ Returns True if the action is relevant to the subgoal, i.e.:
- the action achieves an element of the effects
- the action doesn't delete something that needs to be achieved
- the preconditions are consistent with other subgoals that need to be achieved
@@ -632,7 +636,7 @@ def goal_test(self, subgoal):
def h(self, subgoal):
"""
Computes ignore delete lists heuristic by creating a relaxed version of the original problem (we can do that
- by removing the delete lists from all actions, ie. removing all negative literals from effects) that will be
+ by removing the delete lists from all actions, i.e. removing all negative literals from effects) that will be
easier to solve through GraphPlan and where the length of the solution will serve as a good heuristic.
"""
relaxed_planning_problem = PlanningProblem(initial=self.goal,
@@ -642,12 +646,13 @@ def h(self, subgoal):
try:
return len(linearize(GraphPlan(relaxed_planning_problem).execute()))
except:
- return float('inf')
+ return inf
def CSPlan(planning_problem, solution_length, CSP_solver=ac_search_solver, arc_heuristic=sat_up):
"""
- Planning as Constraint Satisfaction Problem [Section 10.4.3]
+ [Section 10.4.3]
+ Planning as Constraint Satisfaction Problem
"""
def st(var, stage):
@@ -720,7 +725,8 @@ def eq_if_not_in(x1, a, x2):
def SATPlan(planning_problem, solution_length, SAT_solver=cdcl_satisfiable):
"""
- Planning as Boolean satisfiability [Section 10.4.1]
+ [Section 10.4.1]
+ Planning as Boolean satisfiability
"""
def expand_transitions(state, actions):
@@ -1296,7 +1302,9 @@ def toposort(self, graph):
if not ordered:
break
yield ordered
- graph = {element: (dependency - ordered) for element, dependency in graph.items() if element not in ordered}
+ graph = {element: (dependency - ordered)
+ for element, dependency in graph.items()
+ if element not in ordered}
if len(graph) != 0:
raise ValueError('The graph is not acyclic and cannot be linearly ordered')
@@ -1414,8 +1422,7 @@ class HLA(Action):
"""
unique_group = 1
- def __init__(self, action, precond=None, effect=None, duration=0,
- consume=None, use=None):
+ def __init__(self, action, precond=None, effect=None, duration=0, consume=None, use=None):
"""
As opposed to actions, to define HLA, we have added constraints.
duration holds the amount of time required to execute the task
@@ -1437,7 +1444,6 @@ def do_action(self, job_order, available_resources, kb, args):
An HLA based version of act - along with knowledge base updation, it handles
resource checks, and ensures the actions are executed in the correct order.
"""
- # print(self.name)
if not self.has_usable_resource(available_resources):
raise Exception('Not enough usable resources to execute {}'.format(self.name))
if not self.has_consumable_resource(available_resources):
@@ -1517,10 +1523,10 @@ def act(self, action):
raise Exception("Action '{}' not found".format(action.name))
self.initial = list_action.do_action(self.jobs, self.resources, self.initial, args).clauses
- def refinements(hla, library): # refinements may be (multiple) HLA themselves ...
+ def refinements(self, library): # refinements may be (multiple) HLA themselves ...
"""
- state is a Problem, containing the current state kb
- library is a dictionary containing details for every possible refinement. eg:
+ State is a Problem, containing the current state kb library is a
+ dictionary containing details for every possible refinement. e.g.:
{
'HLA': [
'Go(Home, SFO)',
@@ -1550,10 +1556,9 @@ def refinements(hla, library): # refinements may be (multiple) HLA themselves .
['At(SFOLongTermParking) & ~At(Home)'],
['At(SFO) & ~At(SFOLongTermParking)'],
['At(SFO) & ~At(Home)']
- ]
- }
+ ]}
"""
- indices = [i for i, x in enumerate(library['HLA']) if expr(x).op == hla.name]
+ indices = [i for i, x in enumerate(library['HLA']) if expr(x).op == self.name]
for i in indices:
actions = []
for j in range(len(library['steps'][i])):
@@ -1564,14 +1569,15 @@ def refinements(hla, library): # refinements may be (multiple) HLA themselves .
actions.append(HLA(library['steps'][i][j], precond, effect))
yield actions
- def hierarchical_search(problem, hierarchy):
+ def hierarchical_search(self, hierarchy):
"""
- [Figure 11.5] 'Hierarchical Search, a Breadth First Search implementation of Hierarchical
+ [Figure 11.5]
+ 'Hierarchical Search, a Breadth First Search implementation of Hierarchical
Forward Planning Search'
The problem is a real-world problem defined by the problem class, and the hierarchy is
a dictionary of HLA - refinements (see refinements generator for details)
"""
- act = Node(problem.initial, None, [problem.actions[0]])
+ act = Node(self.initial, None, [self.actions[0]])
frontier = deque()
frontier.append(act)
while True:
@@ -1581,8 +1587,8 @@ def hierarchical_search(problem, hierarchy):
# finds the first non primitive hla in plan actions
(hla, index) = RealWorldPlanningProblem.find_hla(plan, hierarchy)
prefix = plan.action[:index]
- outcome = RealWorldPlanningProblem(RealWorldPlanningProblem.result(problem.initial, prefix), problem.goals,
- problem.actions)
+ outcome = RealWorldPlanningProblem(
+ RealWorldPlanningProblem.result(self.initial, prefix), self.goals, self.actions)
suffix = plan.action[index + 1:]
if not hla: # hla is None and plan is primitive
if outcome.goal_test():
@@ -1598,52 +1604,54 @@ def result(state, actions):
state = a(state, a.args).clauses
return state
- def angelic_search(problem, hierarchy, initialPlan):
+ def angelic_search(self, hierarchy, initial_plan):
"""
- [Figure 11.8] A hierarchical planning algorithm that uses angelic semantics to identify and
+ [Figure 11.8]
+ A hierarchical planning algorithm that uses angelic semantics to identify and
commit to high-level plans that work while avoiding high-level plans that don’t.
The predicate MAKING-PROGRESS checks to make sure that we aren’t stuck in an infinite regression
of refinements.
- At top level, call ANGELIC-SEARCH with [Act ] as the initialPlan.
+ At top level, call ANGELIC-SEARCH with [Act] as the initialPlan.
InitialPlan contains a sequence of HLA's with angelic semantics
- The possible effects of an angelic HLA in initialPlan are :
+ The possible effects of an angelic HLA in initialPlan are:
~ : effect remove
$+: effect possibly add
$-: effect possibly remove
$$: possibly add or remove
"""
- frontier = deque(initialPlan)
+ frontier = deque(initial_plan)
while True:
if not frontier:
return None
plan = frontier.popleft() # sequence of HLA/Angelic HLA's
- opt_reachable_set = RealWorldPlanningProblem.reach_opt(problem.initial, plan)
- pes_reachable_set = RealWorldPlanningProblem.reach_pes(problem.initial, plan)
- if problem.intersects_goal(opt_reachable_set):
+ opt_reachable_set = RealWorldPlanningProblem.reach_opt(self.initial, plan)
+ pes_reachable_set = RealWorldPlanningProblem.reach_pes(self.initial, plan)
+ if self.intersects_goal(opt_reachable_set):
if RealWorldPlanningProblem.is_primitive(plan, hierarchy):
return [x for x in plan.action]
- guaranteed = problem.intersects_goal(pes_reachable_set)
- if guaranteed and RealWorldPlanningProblem.making_progress(plan, initialPlan):
+ guaranteed = self.intersects_goal(pes_reachable_set)
+ if guaranteed and RealWorldPlanningProblem.making_progress(plan, initial_plan):
final_state = guaranteed[0] # any element of guaranteed
return RealWorldPlanningProblem.decompose(hierarchy, final_state, pes_reachable_set)
# there should be at least one HLA/Angelic_HLA, otherwise plan would be primitive
hla, index = RealWorldPlanningProblem.find_hla(plan, hierarchy)
prefix = plan.action[:index]
suffix = plan.action[index + 1:]
- outcome = RealWorldPlanningProblem(RealWorldPlanningProblem.result(problem.initial, prefix),
- problem.goals, problem.actions)
+ outcome = RealWorldPlanningProblem(
+ RealWorldPlanningProblem.result(self.initial, prefix), self.goals, self.actions)
for sequence in RealWorldPlanningProblem.refinements(hla, hierarchy): # find refinements
frontier.append(
AngelicNode(outcome.initial, plan, prefix + sequence + suffix, prefix + sequence + suffix))
- def intersects_goal(problem, reachable_set):
+ def intersects_goal(self, reachable_set):
"""
Find the intersection of the reachable states and the goal
"""
- return [y for x in list(reachable_set.keys()) for y in reachable_set[x] if
- all(goal in y for goal in problem.goals)]
+ return [y for x in list(reachable_set.keys())
+ for y in reachable_set[x]
+ if all(goal in y for goal in self.goals)]
def is_primitive(plan, library):
"""
@@ -1706,7 +1714,7 @@ def find_hla(plan, hierarchy):
break
return hla, index
- def making_progress(plan, initialPlan):
+ def making_progress(plan, initial_plan):
"""
Prevents from infinite regression of refinements
@@ -1714,8 +1722,8 @@ def making_progress(plan, initialPlan):
its pessimistic reachable set intersects the goal inside a call to decompose on
the same plan, in the same circumstances)
"""
- for i in range(len(initialPlan)):
- if plan == initialPlan[i]:
+ for i in range(len(initial_plan)):
+ if plan == initial_plan[i]:
return False
return True
@@ -1746,8 +1754,8 @@ def find_previous_state(s_f, reachable_set, i, action):
"""
s_i = reachable_set[i - 1][0]
for state in reachable_set[i - 1]:
- if s_f in [x for x in
- RealWorldPlanningProblem.reach_pes(state, AngelicNode(state, None, [action], [action]))[1]]:
+ if s_f in [x for x in RealWorldPlanningProblem.reach_pes(
+ state, AngelicNode(state, None, [action], [action]))[1]]:
s_i = state
break
return s_i
@@ -1842,9 +1850,7 @@ def go_to_sfo():
['At(SFO) & ~At(Home)'],
['At(SFOLongTermParking) & ~At(Home)'],
['At(SFO) & ~At(SFOLongTermParking)'],
- ['At(SFO) & ~At(Home)']
- ]
- }
+ ['At(SFO) & ~At(Home)']]}
return RealWorldPlanningProblem(initial='At(Home)', goals='At(SFO)', actions=actions), library
@@ -1959,7 +1965,6 @@ def angelic_action(self):
effects[i] = expr(clause.op[w:]) # make changes in the ith part of effects
if n == 3:
effects[i + len(effects) // 3] = expr(clause.op[6:])
- # print('effects', effects)
return [HLA(Expr(self.name, self.args), self.precond, effects[i]) for i in range(len(effects))]
diff --git a/probability.py b/probability.py
index e3fe6cddb..183edfcf8 100644
--- a/probability.py
+++ b/probability.py
@@ -1,11 +1,10 @@
-"""Probability models. (Chapter 13-15)
+"""
+Probability models. (Chapter 13-15)
"""
-from utils import (
- product, argmax, element_wise_product, matrix_multiplication,
- vector_to_diagonal, vector_add, scalar_vector_product, inverse_matrix,
- weighted_sample_with_replacement, isclose, probability, normalize,
- extend)
+from utils import (product, argmax, element_wise_product, matrix_multiplication, vector_to_diagonal, vector_add,
+ scalar_vector_product, inverse_matrix, weighted_sample_with_replacement, isclose, probability,
+ normalize, extend)
from agents import Agent
import random
@@ -18,12 +17,13 @@
def DTAgentProgram(belief_state):
- """A decision-theoretic agent. [Figure 13.1]"""
+ """
+ [Figure 13.1]
+ A decision-theoretic agent."""
def program(percept):
belief_state.observe(program.action, percept)
- program.action = argmax(belief_state.actions(),
- key=belief_state.expected_outcome_utility)
+ program.action = argmax(belief_state.actions(), key=belief_state.expected_outcome_utility)
return program.action
program.action = None
@@ -43,11 +43,11 @@ class ProbDist:
(0.125, 0.375, 0.5)
"""
- def __init__(self, varname='?', freqs=None):
+ def __init__(self, var_name='?', freqs=None):
"""If freqs is given, it is a dictionary of values - frequency pairs,
then ProbDist is normalized."""
self.prob = {}
- self.varname = varname
+ self.var_name = var_name
self.values = []
if freqs:
for (v, p) in freqs.items():
@@ -80,11 +80,10 @@ def normalize(self):
def show_approx(self, numfmt='{:.3g}'):
"""Show the probabilities rounded and sorted by key, for the
sake of portable doctests."""
- return ', '.join([('{}: ' + numfmt).format(v, p)
- for (v, p) in sorted(self.prob.items())])
+ return ', '.join([('{}: ' + numfmt).format(v, p) for (v, p) in sorted(self.prob.items())])
def __repr__(self):
- return "P({})".format(self.varname)
+ return "P({})".format(self.var_name)
class JointProbDist(ProbDist):
@@ -141,8 +140,10 @@ def event_values(event, variables):
def enumerate_joint_ask(X, e, P):
- """Return a probability distribution over the values of the variable X,
- given the {var:val} observations e, in the JointProbDist P. [Section 13.3]
+ """
+ [Section 13.3]
+ Return a probability distribution over the values of the variable X,
+ given the {var:val} observations e, in the JointProbDist P.
>>> P = JointProbDist(['X', 'Y'])
>>> P[0,0] = 0.25; P[0,1] = 0.5; P[1,1] = P[2,1] = 0.125
>>> enumerate_joint_ask('X', dict(Y=1), P).show_approx()
@@ -239,9 +240,11 @@ def get_expected_utility(self, action, evidence):
class InformationGatheringAgent(Agent):
- """A simple information gathering agent. The agent works by repeatedly selecting
+ """
+ [Figure 16.9]
+ A simple information gathering agent. The agent works by repeatedly selecting
the observation with the highest information value, until the cost of the next
- observation is greater than its expected benefit. [Figure 16.9]"""
+ observation is greater than its expected benefit."""
def __init__(self, decnet, infer, initial_evidence=None):
"""decnet: a decision network
@@ -381,16 +384,17 @@ def __repr__(self):
('Alarm', 'Burglary Earthquake',
{(T, T): 0.95, (T, F): 0.94, (F, T): 0.29, (F, F): 0.001}),
('JohnCalls', 'Alarm', {T: 0.90, F: 0.05}),
- ('MaryCalls', 'Alarm', {T: 0.70, F: 0.01})
-])
+ ('MaryCalls', 'Alarm', {T: 0.70, F: 0.01})])
# ______________________________________________________________________________
def enumeration_ask(X, e, bn):
- """Return the conditional probability distribution of variable X
- given evidence e, from BayesNet bn. [Figure 14.9]
+ """
+ [Figure 14.9]
+ Return the conditional probability distribution of variable X
+ given evidence e, from BayesNet bn.
>>> enumeration_ask('Burglary', dict(JohnCalls=T, MaryCalls=T), burglary
... ).show_approx()
'False: 0.716, True: 0.284'"""
@@ -421,7 +425,9 @@ def enumerate_all(variables, e, bn):
def elimination_ask(X, e, bn):
- """Compute bn's P(X|e) by variable elimination. [Figure 14.11]
+ """
+ [Figure 14.11]
+ Compute bn's P(X|e) by variable elimination.
>>> elimination_ask('Burglary', dict(JohnCalls=T, MaryCalls=T), burglary
... ).show_approx()
'False: 0.716, True: 0.284'"""
@@ -473,23 +479,20 @@ def __init__(self, variables, cpt):
def pointwise_product(self, other, bn):
"""Multiply two factors, combining their variables."""
variables = list(set(self.variables) | set(other.variables))
- cpt = {event_values(e, variables): self.p(e) * other.p(e)
- for e in all_events(variables, bn, {})}
+ cpt = {event_values(e, variables): self.p(e) * other.p(e) for e in all_events(variables, bn, {})}
return Factor(variables, cpt)
def sum_out(self, var, bn):
"""Make a factor eliminating var by summing over its values."""
variables = [X for X in self.variables if X != var]
- cpt = {event_values(e, variables): sum(self.p(extend(e, var, val))
- for val in bn.variable_values(var))
+ cpt = {event_values(e, variables): sum(self.p(extend(e, var, val)) for val in bn.variable_values(var))
for e in all_events(variables, bn, {})}
return Factor(variables, cpt)
def normalize(self):
"""Return my probabilities; must be down to one variable."""
assert len(self.variables) == 1
- return ProbDist(self.variables[0],
- {k: v for ((k,), v) in self.cpt.items()})
+ return ProbDist(self.variables[0], {k: v for ((k,), v) in self.cpt.items()})
def p(self, e):
"""Look up my value tabulated for e."""
@@ -524,8 +527,10 @@ def all_events(variables, bn, e):
def prior_sample(bn):
- """Randomly sample from bn's full joint distribution. The result
- is a {variable: value} dict. [Figure 14.13]"""
+ """
+ [Figure 14.13]
+ Randomly sample from bn's full joint distribution. The result
+ is a {variable: value} dict."""
event = {}
for node in bn.nodes:
event[node.variable] = node.sample(event)
@@ -555,16 +560,17 @@ def rejection_sampling(X, e, bn, N=10000):
def consistent_with(event, evidence):
"""Is event consistent with the given evidence?"""
- return all(evidence.get(k, v) == v
- for k, v in event.items())
+ return all(evidence.get(k, v) == v for k, v in event.items())
# _________________________________________________________________________
def likelihood_weighting(X, e, bn, N=10000):
- """Estimate the probability distribution of variable X given
- evidence e in BayesNet bn. [Figure 14.15]
+ """
+ [Figure 14.15]
+ Estimate the probability distribution of variable X given
+ evidence e in BayesNet bn.
>>> random.seed(1017)
>>> likelihood_weighting('Burglary', dict(JohnCalls=T, MaryCalls=T),
... burglary, 10000).show_approx()
@@ -619,9 +625,8 @@ def markov_blanket_sample(X, e, bn):
Q = ProbDist(X)
for xi in bn.variable_values(X):
ei = extend(e, X, xi)
- # [Equation 14.12:]
- Q[xi] = Xnode.p(xi, e) * product(Yj.p(ei[Yj.variable], ei)
- for Yj in Xnode.children)
+ # [Equation 14.12]
+ Q[xi] = Xnode.p(xi, e) * product(Yj.p(ei[Yj.variable], ei) for Yj in Xnode.children)
# (assuming a Boolean variable here)
return probability(Q.normalize()[True])
@@ -661,7 +666,8 @@ def backward(HMM, b, ev):
def forward_backward(HMM, ev):
- """[Figure 15.4]
+ """
+ [Figure 15.4]
Forward-Backward algorithm for smoothing. Computes posterior probabilities
of a sequence of states given a sequence of observations."""
t = len(ev)
@@ -687,9 +693,10 @@ def forward_backward(HMM, ev):
def viterbi(HMM, ev):
- """[Equation 15.11]
- Viterbi algorithm to find the most likely sequence. Computes the best path and the corresponding probabilities,
- given an HMM model and a sequence of observations."""
+ """
+ [Equation 15.11]
+ Viterbi algorithm to find the most likely sequence. Computes the best path and the
+ corresponding probabilities, given an HMM model and a sequence of observations."""
t = len(ev)
ev = ev.copy()
ev.insert(0, None)
@@ -713,8 +720,8 @@ def viterbi(HMM, ev):
# most likely sequence
ml_path = [True] * (len(ev) - 1)
- # the construction of the most likely sequence starts in the final state with the largest probability,
- # and runs backwards; the algorithm needs to store for each xt its predecessor xt-1 maximizing its probability
+ # the construction of the most likely sequence starts in the final state with the largest probability, and
+ # runs backwards; the algorithm needs to store for each xt its predecessor xt-1 maximizing its probability
i_max = np.argmax(m[-1])
for i in range(t - 1, -1, -1):
@@ -730,7 +737,8 @@ def viterbi(HMM, ev):
def fixed_lag_smoothing(e_t, HMM, d, ev, t):
- """[Figure 15.6]
+ """
+ [Figure 15.6]
Smoothing algorithm with a fixed time lag of 'd' steps.
Online algorithm that outputs the new smoothed estimate if observation
for new time step is given."""
@@ -842,7 +850,9 @@ def ray_cast(self, sensor_num, kin_state):
def monte_carlo_localization(a, z, N, P_motion_sample, P_sensor, m, S=None):
- """Monte Carlo localization algorithm from Fig 25.9"""
+ """
+ [Figure 25.9]
+ Monte Carlo localization algorithm"""
def ray_cast(sensor_num, kin_state, m):
return m.ray_cast(sensor_num, kin_state)
diff --git a/probability4e.py b/probability4e.py
index dca88d4ad..7d464c62a 100644
--- a/probability4e.py
+++ b/probability4e.py
@@ -1,7 +1,6 @@
-"""Probability models.
-"""
+"""Probability models."""
-from utils import product, argmax, isclose, probability, extend
+from utils4e import product, argmax, isclose, probability, extend
from math import sqrt, pi, exp
import copy
import random
diff --git a/reinforcement_learning4e.py b/reinforcement_learning4e.py
index 86c268544..44fda5c87 100644
--- a/reinforcement_learning4e.py
+++ b/reinforcement_learning4e.py
@@ -1,7 +1,7 @@
"""Reinforcement Learning (Chapter 21)"""
from collections import defaultdict
-from utils import argmax
+from utils4e import argmax
from mdp import MDP, policy_evaluation
import random
diff --git a/requirements.txt b/requirements.txt
index 5a6603dd8..bf019e803 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -1,3 +1,5 @@
+ipywidgets
+scipy
pytest
sortedcontainers
networkx
diff --git a/search.py b/search.py
index 2491dc6e5..87f6b86e3 100644
--- a/search.py
+++ b/search.py
@@ -1,8 +1,10 @@
-"""Search (Chapters 3-4)
+"""
+Search (Chapters 3-4)
The way to use this code is to subclass Problem to create a class of problems,
then create problem instances and solve them with calls to the various search
-functions."""
+functions.
+"""
import bisect
import math
@@ -10,19 +12,14 @@
import sys
from collections import deque
-from utils import (
- is_in, argmin, argmax, argmax_random_tie, probability, weighted_sampler,
- memoize, print_table, open_data, PriorityQueue, name,
- distance, vector_add
-)
-
-infinity = float('inf')
+from utils import (is_in, argmin, argmax, argmax_random_tie, probability, weighted_sampler, memoize,
+ print_table, open_data, PriorityQueue, name, distance, vector_add, inf)
# ______________________________________________________________________________
-class Problem(object):
+class Problem:
"""The abstract class for a formal problem. You should subclass
this and implement the methods actions and result, and possibly
__init__, goal_test, and path_cost. Then you will create instances
@@ -109,9 +106,7 @@ def expand(self, problem):
def child_node(self, problem, action):
"""[Figure 3.10]"""
next_state = problem.result(self.state, action)
- next_node = Node(next_state, self, action,
- problem.path_cost(self.path_cost, self.state,
- action, next_state))
+ next_node = Node(next_state, self, action, problem.path_cost(self.path_cost, self.state, action, next_state))
return next_node
def solution(self):
@@ -219,6 +214,7 @@ def depth_first_graph_search(problem):
Does not get trapped by loops.
If two paths reach a state, only use the first one. [Figure 3.7]"""
frontier = [(Node(problem.initial))] # Stack
+
explored = set()
while frontier:
node = frontier.pop()
@@ -226,8 +222,7 @@ def depth_first_graph_search(problem):
return node
explored.add(node.state)
frontier.extend(child for child in node.expand(problem)
- if child.state not in explored and
- child not in frontier)
+ if child.state not in explored and child not in frontier)
return None
@@ -253,7 +248,7 @@ def breadth_first_graph_search(problem):
return None
-def best_first_graph_search(problem, f):
+def best_first_graph_search(problem, f, display=False):
"""Search the nodes with the lowest f scores first.
You specify the function f(node) that you want to minimize; for example,
if f is a heuristic estimate to the goal, then we have greedy best
@@ -269,6 +264,8 @@ def best_first_graph_search(problem, f):
while frontier:
node = frontier.pop()
if problem.goal_test(node.state):
+ if display:
+ print(len(explored), "paths have been expanded and", len(frontier), "paths remain in the frontier")
return node
explored.add(node.state)
for child in node.expand(problem):
@@ -281,9 +278,9 @@ def best_first_graph_search(problem, f):
return None
-def uniform_cost_search(problem):
+def uniform_cost_search(problem, display=False):
"""[Figure 3.14]"""
- return best_first_graph_search(problem, lambda node: node.path_cost)
+ return best_first_graph_search(problem, lambda node: node.path_cost, display)
def depth_limited_search(problem, limit=50):
@@ -325,7 +322,7 @@ def bidirectional_search(problem):
gF, gB = {problem.initial: 0}, {problem.goal: 0}
openF, openB = [problem.initial], [problem.goal]
closedF, closedB = [], []
- U = infinity
+ U = inf
def extend(U, open_dir, open_other, g_dir, g_other, closed_dir):
"""Extend search in given direction"""
@@ -351,7 +348,7 @@ def extend(U, open_dir, open_other, g_dir, g_other, closed_dir):
def find_min(open_dir, g):
"""Finds minimum priority, g and f values in open_dir"""
- m, m_f = infinity, infinity
+ m, m_f = inf, inf
for n in open_dir:
f = g[n] + problem.h(n)
pr = max(f, 2 * g[n])
@@ -363,7 +360,7 @@ def find_min(open_dir, g):
def find_key(pr_min, open_dir, g):
"""Finds key in open_dir with value equal to pr_min
and minimum g value."""
- m = infinity
+ m = inf
state = -1
for n in open_dir:
pr = max(g[n] + problem.h(n), 2 * g[n])
@@ -389,7 +386,7 @@ def find_key(pr_min, open_dir, g):
# Extend backward
U, openB, closedB, gB = extend(U, openB, openF, gB, gF, closedB)
- return infinity
+ return inf
# ______________________________________________________________________________
@@ -402,21 +399,21 @@ def find_key(pr_min, open_dir, g):
# Greedy best-first search is accomplished by specifying f(n) = h(n).
-def astar_search(problem, h=None):
+def astar_search(problem, h=None, display=False):
"""A* search is best-first graph search with f(n) = g(n)+h(n).
You need to specify the h function when you call astar_search, or
else in your Problem subclass."""
h = memoize(h or problem.h, 'h')
- return best_first_graph_search(problem, lambda n: n.path_cost + h(n))
+ return best_first_graph_search(problem, lambda n: n.path_cost + h(n), display)
# ______________________________________________________________________________
# A* heuristics
class EightPuzzle(Problem):
- """ The problem of sliding tiles numbered from 1 to 8 on a 3x3 board,
- where one of the squares is a blank. A state is represented as a tuple of length 9,
- where element at index i represents the tile number at index i (0 if it's an empty square) """
+ """ The problem of sliding tiles numbered from 1 to 8 on a 3x3 board, where one of the
+ squares is a blank. A state is represented as a tuple of length 9, where element at
+ index i represents the tile number at index i (0 if it's an empty square) """
def __init__(self, initial, goal=(1, 2, 3, 4, 5, 6, 7, 8, 0)):
""" Define goal state and initialize a problem """
@@ -602,7 +599,7 @@ def RBFS(problem, node, flimit):
return node, 0 # (The second value is immaterial)
successors = node.expand(problem)
if len(successors) == 0:
- return None, infinity
+ return None, inf
for s in successors:
s.f = max(s.path_cost + h(s), node.f)
while True:
@@ -614,14 +611,14 @@ def RBFS(problem, node, flimit):
if len(successors) > 1:
alternative = successors[1].f
else:
- alternative = infinity
+ alternative = inf
result, best.f = RBFS(problem, best, min(flimit, alternative))
if result is not None:
return result, best.f
node = Node(problem.initial)
node.f = h(node)
- result, bestf = RBFS(problem, node, infinity)
+ result, bestf = RBFS(problem, node, inf)
return result
@@ -1072,7 +1069,7 @@ def RandomGraph(nodes=list(range(10)), min_links=2, width=400, height=300,
def distance_to_node(n):
if n is node or g.get(node, n):
- return infinity
+ return inf
return distance(g.locations[n], here)
neighbor = argmin(nodes, key=distance_to_node)
@@ -1180,11 +1177,11 @@ def result(self, state, action):
return action
def path_cost(self, cost_so_far, A, action, B):
- return cost_so_far + (self.graph.get(A, B) or infinity)
+ return cost_so_far + (self.graph.get(A, B) or inf)
def find_min_edge(self):
"""Find minimum value of edges."""
- m = infinity
+ m = inf
for d in self.graph.graph_dict.values():
local_min = min(d.values())
m = min(m, local_min)
@@ -1200,7 +1197,7 @@ def h(self, node):
return int(distance(locs[node.state], locs[self.goal]))
else:
- return infinity
+ return inf
class GraphProblemStochastic(GraphProblem):
diff --git a/tests/test_csp.py b/tests/test_csp.py
index 6aafa81c8..553880a40 100644
--- a/tests/test_csp.py
+++ b/tests/test_csp.py
@@ -176,7 +176,8 @@ def test_revise():
Xj = 'B'
removals = []
- assert not revise(csp, Xi, Xj, removals)
+ consistency, _ = revise(csp, Xi, Xj, removals)
+ assert not consistency
assert len(removals) == 0
domains = {'A': [0, 1, 2, 3, 4], 'B': [0, 1, 2, 3, 4]}
@@ -195,7 +196,8 @@ def test_AC3():
csp = CSP(variables=None, domains=domains, neighbors=neighbors, constraints=constraints)
- assert not AC3(csp, removals=removals)
+ consistency, _ = AC3(csp, removals=removals)
+ assert not consistency
constraints = lambda X, x, Y, y: x % 2 == 0 and x + y == 4
removals = []
@@ -221,7 +223,8 @@ def test_AC3b():
csp = CSP(variables=None, domains=domains, neighbors=neighbors, constraints=constraints)
- assert not AC3b(csp, removals=removals)
+ consistency, _ = AC3b(csp, removals=removals)
+ assert not consistency
constraints = lambda X, x, Y, y: x % 2 == 0 and x + y == 4
removals = []
@@ -247,7 +250,8 @@ def test_AC4():
csp = CSP(variables=None, domains=domains, neighbors=neighbors, constraints=constraints)
- assert not AC4(csp, removals=removals)
+ consistency, _ = AC4(csp, removals=removals)
+ assert not consistency
constraints = lambda X, x, Y, y: x % 2 == 0 and x + y == 4
removals = []
@@ -492,8 +496,8 @@ def test_ac_solver():
'four_across': 'car'}
assert ac_solver(two_two_four) == {'T': 7, 'F': 1, 'W': 6, 'O': 5, 'U': 3, 'R': 0, 'C1': 1, 'C2': 1, 'C3': 1} or \
{'T': 9, 'F': 1, 'W': 2, 'O': 8, 'U': 5, 'R': 6, 'C1': 1, 'C2': 0, 'C3': 1}
- assert ac_solver(send_more_money) == {'S': 9, 'M': 1, 'E': 5, 'N': 6, 'D': 7, 'O': 0, 'R': 8, 'Y': 2,
- 'C1': 1, 'C2': 1, 'C3': 0, 'C4': 1}
+ assert ac_solver(send_more_money) == \
+ {'S': 9, 'M': 1, 'E': 5, 'N': 6, 'D': 7, 'O': 0, 'R': 8, 'Y': 2, 'C1': 1, 'C2': 1, 'C3': 0, 'C4': 1}
def test_ac_search_solver():
@@ -614,11 +618,13 @@ def test_mac():
assignment = {'A': 1}
csp = CSP(variables=None, domains=domains, neighbors=neighbors, constraints=constraints)
- assert not mac(csp, var, value, assignment, None)
+ consistency, _ = mac(csp, var, value, assignment, None)
+ assert not consistency
constraints = lambda X, x, Y, y: x % 2 != 0 and (x + y) == 6 and y % 2 != 0
csp = CSP(variables=None, domains=domains, neighbors=neighbors, constraints=constraints)
- assert mac(csp, var, value, assignment, None)
+ _, consistency = mac(csp, var, value, assignment, None)
+ assert consistency
def test_queen_constraint():
diff --git a/tests/test_knowledge.py b/tests/test_knowledge.py
index 6b65bd87f..556637652 100644
--- a/tests/test_knowledge.py
+++ b/tests/test_knowledge.py
@@ -103,38 +103,38 @@ def test_minimal_consistent_det():
A, B, C, D, E, F, G, H, I, x, y, z = map(expr, 'ABCDEFGHIxyz')
# knowledge base containing family relations
-small_family = FOIL_container([expr("Mother(Anne, Peter)"),
- expr("Mother(Anne, Zara)"),
- expr("Mother(Sarah, Beatrice)"),
- expr("Mother(Sarah, Eugenie)"),
- expr("Father(Mark, Peter)"),
- expr("Father(Mark, Zara)"),
- expr("Father(Andrew, Beatrice)"),
- expr("Father(Andrew, Eugenie)"),
- expr("Father(Philip, Anne)"),
- expr("Father(Philip, Andrew)"),
- expr("Mother(Elizabeth, Anne)"),
- expr("Mother(Elizabeth, Andrew)"),
- expr("Male(Philip)"),
- expr("Male(Mark)"),
- expr("Male(Andrew)"),
- expr("Male(Peter)"),
- expr("Female(Elizabeth)"),
- expr("Female(Anne)"),
- expr("Female(Sarah)"),
- expr("Female(Zara)"),
- expr("Female(Beatrice)"),
- expr("Female(Eugenie)")])
-
-smaller_family = FOIL_container([expr("Mother(Anne, Peter)"),
- expr("Father(Mark, Peter)"),
- expr("Father(Philip, Anne)"),
- expr("Mother(Elizabeth, Anne)"),
- expr("Male(Philip)"),
- expr("Male(Mark)"),
- expr("Male(Peter)"),
- expr("Female(Elizabeth)"),
- expr("Female(Anne)")])
+small_family = FOILContainer([expr("Mother(Anne, Peter)"),
+ expr("Mother(Anne, Zara)"),
+ expr("Mother(Sarah, Beatrice)"),
+ expr("Mother(Sarah, Eugenie)"),
+ expr("Father(Mark, Peter)"),
+ expr("Father(Mark, Zara)"),
+ expr("Father(Andrew, Beatrice)"),
+ expr("Father(Andrew, Eugenie)"),
+ expr("Father(Philip, Anne)"),
+ expr("Father(Philip, Andrew)"),
+ expr("Mother(Elizabeth, Anne)"),
+ expr("Mother(Elizabeth, Andrew)"),
+ expr("Male(Philip)"),
+ expr("Male(Mark)"),
+ expr("Male(Andrew)"),
+ expr("Male(Peter)"),
+ expr("Female(Elizabeth)"),
+ expr("Female(Anne)"),
+ expr("Female(Sarah)"),
+ expr("Female(Zara)"),
+ expr("Female(Beatrice)"),
+ expr("Female(Eugenie)")])
+
+smaller_family = FOILContainer([expr("Mother(Anne, Peter)"),
+ expr("Father(Mark, Peter)"),
+ expr("Father(Philip, Anne)"),
+ expr("Mother(Elizabeth, Anne)"),
+ expr("Male(Philip)"),
+ expr("Male(Mark)"),
+ expr("Male(Peter)"),
+ expr("Female(Elizabeth)"),
+ expr("Female(Anne)")])
# target relation
target = expr('Parent(x, y)')
diff --git a/tests/test_logic.py b/tests/test_logic.py
index a680951e3..c05b29ec1 100644
--- a/tests/test_logic.py
+++ b/tests/test_logic.py
@@ -183,13 +183,28 @@ def test_unify():
assert unify(expr('American(x) & Weapon(B)'), expr('American(A) & Weapon(y)')) == {x: A, y: B}
assert unify(expr('P(F(x,z), G(u, z))'), expr('P(F(y,a), y)')) == {x: G(u, a), z: a, y: G(u, a)}
- # test for https://github.com/aimacode/aima-python/issues/1053
+ # tests for https://github.com/aimacode/aima-python/issues/1053
# unify(expr('P(A, x, F(G(y)))'), expr('P(z, F(z), F(u))'))
# must return {z: A, x: F(A), u: G(y)} and not {z: A, x: F(z), u: G(y)}
assert unify(expr('P(A, x, F(G(y)))'), expr('P(z, F(z), F(u))')) == {z: A, x: F(A), u: G(y)}
assert unify(expr('P(x, A, F(G(y)))'), expr('P(F(z), z, F(u))')) == {x: F(A), z: A, u: G(y)}
+def test_unify_mm():
+ assert unify_mm(x, x) == {}
+ assert unify_mm(x, 3) == {x: 3}
+ assert unify_mm(x & 4 & y, 6 & y & 4) == {x: 6, y: 4}
+ assert unify_mm(expr('A(x)'), expr('A(B)')) == {x: B}
+ assert unify_mm(expr('American(x) & Weapon(B)'), expr('American(A) & Weapon(y)')) == {x: A, y: B}
+ assert unify_mm(expr('P(F(x,z), G(u, z))'), expr('P(F(y,a), y)')) == {x: G(u, a), z: a, y: G(u, a)}
+
+ # tests for https://github.com/aimacode/aima-python/issues/1053
+ # unify(expr('P(A, x, F(G(y)))'), expr('P(z, F(z), F(u))'))
+ # must return {z: A, x: F(A), u: G(y)} and not {z: A, x: F(z), u: G(y)}
+ assert unify_mm(expr('P(A, x, F(G(y)))'), expr('P(z, F(z), F(u))')) == {z: A, x: F(A), u: G(y)}
+ assert unify_mm(expr('P(x, A, F(G(y)))'), expr('P(F(z), z, F(u))')) == {x: F(A), z: A, u: G(y)}
+
+
def test_pl_fc_entails():
assert pl_fc_entails(horn_clauses_KB, expr('Q'))
assert pl_fc_entails(definite_clauses_KB, expr('G'))
diff --git a/tests/test_perception4e.py b/tests/test_perception4e.py
index b6105e25e..ee5f12fd9 100644
--- a/tests/test_perception4e.py
+++ b/tests/test_perception4e.py
@@ -75,9 +75,11 @@ def test_ROIPoolingLayer():
feature_map = np.ones(feature_maps_shape, dtype='float32')
feature_map[200 - 1, 100 - 3, 0] = 50
roiss = np.asarray([[0.5, 0.2, 0.7, 0.4], [0.0, 0.0, 1.0, 1.0]])
- assert pool_rois(feature_map, roiss, 3, 7)[0].tolist() == [[1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1],
+ assert pool_rois(feature_map, roiss, 3, 7)[0].tolist() == [[1, 1, 1, 1, 1, 1, 1],
+ [1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 1]]
- assert pool_rois(feature_map, roiss, 3, 7)[1].tolist() == [[1, 1, 1, 1, 1, 1, 1], [1, 1, 1, 1, 1, 1, 1],
+ assert pool_rois(feature_map, roiss, 3, 7)[1].tolist() == [[1, 1, 1, 1, 1, 1, 1],
+ [1, 1, 1, 1, 1, 1, 1],
[1, 1, 1, 1, 1, 1, 50]]
diff --git a/tests/test_planning.py b/tests/test_planning.py
index cb51dc090..103402481 100644
--- a/tests/test_planning.py
+++ b/tests/test_planning.py
@@ -560,8 +560,7 @@ def test_job_shop_problem():
['At(MetroStop)'], ['At(Home) & Have(Cash)']],
'effect': [['At(SFO) & ~At(Home)'], ['At(SFO) & ~At(Home) & ~Have(Cash)'], ['At(MetroStop) & ~At(Home)'],
['At(SFO) & ~At(MetroStop)'], ['At(SFO) & ~At(MetroStop)'], ['At(SFO) & ~At(MetroStop)'],
- ['At(SFO) & ~At(MetroStop)'], ['At(SFO) & ~At(Home) & ~Have(Cash)']]
-}
+ ['At(SFO) & ~At(MetroStop)'], ['At(SFO) & ~At(Home) & ~Have(Cash)']]}
# HLA's
go_SFO = HLA('Go(Home,SFO)', precond='At(Home)', effect='At(SFO) & ~At(Home)')
diff --git a/tests/test_probability.py b/tests/test_probability.py
index b38052894..8def79c68 100644
--- a/tests/test_probability.py
+++ b/tests/test_probability.py
@@ -145,7 +145,7 @@ def test_enumeration_ask():
burglary).show_approx() == 'False: 0.944, True: 0.0561'
-def test_elemination_ask():
+def test_elimination_ask():
assert elimination_ask(
'Burglary', dict(JohnCalls=T, MaryCalls=T),
burglary).show_approx() == 'False: 0.716, True: 0.284'
diff --git a/tests/test_utils.py b/tests/test_utils.py
index 672784bef..6e2bdbcdd 100644
--- a/tests/test_utils.py
+++ b/tests/test_utils.py
@@ -158,9 +158,9 @@ def test_mean_error():
assert mean_error([0, 0.5], [0, -0.5]) == 0.5
-def test_dotproduct():
- assert dotproduct([1, 2, 3], [1000, 100, 10]) == 1230
- assert dotproduct([1, 2, 3], [0, 0, 0]) == 0
+def test_dot_product():
+ assert dot_product([1, 2, 3], [1000, 100, 10]) == 1230
+ assert dot_product([1, 2, 3], [0, 0, 0]) == 0
def test_element_wise_product():
@@ -202,8 +202,7 @@ def test_scalar_vector_product():
def test_scalar_matrix_product():
- assert rounder(scalar_matrix_product(-5, [[1, 2], [3, 4], [0, 6]])) == [[-5, -10], [-15, -20],
- [0, -30]]
+ assert rounder(scalar_matrix_product(-5, [[1, 2], [3, 4], [0, 6]])) == [[-5, -10], [-15, -20], [0, -30]]
assert rounder(scalar_matrix_product(0.2, [[1, 2], [2, 3]])) == [[0.2, 0.4], [0.4, 0.6]]
diff --git a/utils.py b/utils.py
index 75d4547cf..68694532e 100644
--- a/utils.py
+++ b/utils.py
@@ -1,4 +1,4 @@
-"""Provides some utilities widely used by other modules"""
+"""Provides some utilities widely used by other modules."""
import bisect
import collections
@@ -14,6 +14,8 @@
import numpy as np
from itertools import chain, combinations
+inf = float('inf')
+
# ______________________________________________________________________________
# Functions on Sequences and Iterables
@@ -21,8 +23,7 @@
def sequence(iterable):
"""Converts iterable to sequence, if it is not already one."""
- return (iterable if isinstance(iterable, collections.abc.Sequence)
- else tuple([iterable]))
+ return iterable if isinstance(iterable, collections.abc.Sequence) else tuple([iterable])
def remove_all(item, seq):
@@ -141,13 +142,12 @@ def histogram(values, mode=0, bin_function=None):
bins[val] = bins.get(val, 0) + 1
if mode:
- return sorted(list(bins.items()), key=lambda x: (x[1], x[0]),
- reverse=True)
+ return sorted(list(bins.items()), key=lambda x: (x[1], x[0]), reverse=True)
else:
return sorted(bins.items())
-def dotproduct(X, Y):
+def dot_product(X, Y):
"""Return the sum of the element-wise product of vectors X and Y."""
return sum(x * y for x, y in zip(X, Y))
@@ -163,16 +163,12 @@ def matrix_multiplication(X_M, *Y_M):
def _mat_mult(X_M, Y_M):
"""Return a matrix as a matrix-multiplication of two matrices X_M and Y_M
- >>> matrix_multiplication([[1, 2, 3],
- [2, 3, 4]],
- [[3, 4],
- [1, 2],
- [1, 0]])
+ >>> matrix_multiplication([[1, 2, 3], [2, 3, 4]], [[3, 4], [1, 2], [1, 0]])
[[8, 8],[13, 14]]
"""
assert len(X_M[0]) == len(Y_M)
- result = [[0 for i in range(len(Y_M[0]))] for j in range(len(X_M))]
+ result = [[0 for i in range(len(Y_M[0]))] for _ in range(len(X_M))]
for i in range(len(X_M)):
for j in range(len(Y_M[0])):
for k in range(len(Y_M)):
@@ -189,7 +185,7 @@ def _mat_mult(X_M, Y_M):
def vector_to_diagonal(v):
"""Converts a vector to a diagonal matrix with vector elements
as the diagonal elements of the matrix"""
- diag_matrix = [[0 for i in range(len(v))] for j in range(len(v))]
+ diag_matrix = [[0 for i in range(len(v))] for _ in range(len(v))]
for i in range(len(v)):
diag_matrix[i][i] = v[i]
@@ -218,7 +214,6 @@ def inverse_matrix(X):
det = X[0][0] * X[1][1] - X[0][1] * X[1][0]
assert det != 0
inv_mat = scalar_matrix_product(1.0 / det, [[X[1][1], -X[0][1]], [-X[1][0], X[0][0]]])
-
return inv_mat
@@ -232,7 +227,6 @@ def weighted_sample_with_replacement(n, seq, weights):
probability of each element in proportion to its corresponding
weight."""
sample = weighted_sampler(seq, weights)
-
return [sample() for _ in range(n)]
@@ -241,13 +235,12 @@ def weighted_sampler(seq, weights):
totals = []
for w in weights:
totals.append(w + totals[-1] if totals else w)
-
return lambda: seq[bisect.bisect(totals, random.uniform(0, totals[-1]))]
def weighted_choice(choices):
"""A weighted version of random.choice"""
- # NOTE: Shoule be replaced by random.choices if we port to Python 3.6
+ # NOTE: should be replaced by random.choices if we port to Python 3.6
total = sum(w for _, w in choices)
r = random.uniform(0, total)
@@ -268,8 +261,7 @@ def rounder(numbers, d=4):
def num_or_str(x): # TODO: rename as `atom`
- """The argument is a string; convert to a number if
- possible, or strip it."""
+ """The argument is a string; convert to a number if possible, or strip it."""
try:
return int(x)
except ValueError:
@@ -318,7 +310,7 @@ def normalize(dist):
total = sum(dist.values())
for key in dist:
dist[key] = dist[key] / total
- assert 0 <= dist[key] <= 1, "Probabilities must be between 0 and 1."
+ assert 0 <= dist[key] <= 1 # Probabilities must be between 0 and 1
return dist
total = sum(dist)
return [(n / total) for n in dist]
@@ -355,17 +347,11 @@ def relu_derivative(value):
def elu(x, alpha=0.01):
- if x > 0:
- return x
- else:
- return alpha * (math.exp(x) - 1)
+ return x if x > 0 else alpha * (math.exp(x) - 1)
def elu_derivative(value, alpha=0.01):
- if value > 0:
- return 1
- else:
- return alpha * math.exp(value)
+ return 1 if value > 0 else alpha * math.exp(value)
def tanh(x):
@@ -373,21 +359,15 @@ def tanh(x):
def tanh_derivative(value):
- return (1 - (value ** 2))
+ return 1 - (value ** 2)
def leaky_relu(x, alpha=0.01):
- if x > 0:
- return x
- else:
- return alpha * x
+ return x if x > 0 else alpha * x
def leaky_relu_derivative(value, alpha=0.01):
- if value > 0:
- return 1
- else:
- return alpha
+ return 1 if value > 0 else alpha
def relu(x):
@@ -395,10 +375,7 @@ def relu(x):
def relu_derivative(value):
- if value > 0:
- return 1
- else:
- return 0
+ return 1 if value > 0 else 0
def step(x):
@@ -437,10 +414,10 @@ def remove_component(X):
X_m = X[:m]
X_n = X[m:]
for eivec in eivec_m:
- coeff = dotproduct(X_m, eivec)
+ coeff = dot_product(X_m, eivec)
X_m = [x1 - coeff * x2 for x1, x2 in zip(X_m, eivec)]
for eivec in eivec_n:
- coeff = dotproduct(X_n, eivec)
+ coeff = dot_product(X_n, eivec)
X_n = [x1 - coeff * x2 for x1, x2 in zip(X_n, eivec)]
return X_m + X_n
@@ -527,7 +504,7 @@ def vector_clip(vector, lowest, highest):
# ______________________________________________________________________________
# Misc Functions
-class injection():
+class injection:
"""Dependency injection of temporary values for global functions/classes/etc.
E.g., `with injection(DataBase=MockDataBase): ...`"""
@@ -819,10 +796,7 @@ def expr(x):
>>> expr('P & Q ==> Q')
((P & Q) ==> Q)
"""
- if isinstance(x, str):
- return eval(expr_handle_infix_ops(x), defaultkeydict(Symbol))
- else:
- return x
+ return eval(expr_handle_infix_ops(x), defaultkeydict(Symbol)) if isinstance(x, str) else x
infix_ops = '==> <== <=>'.split()
@@ -873,7 +847,6 @@ class PriorityQueue:
def __init__(self, order='min', f=lambda x: x):
self.heap = []
-
if order == 'min':
self.f = f
elif order == 'max': # now item with max f(x)
@@ -923,22 +896,6 @@ def __delitem__(self, key):
heapq.heapify(self.heap)
-# ______________________________________________________________________________
-# Monte Carlo tree node and ucb function
-class MCT_Node:
- """Node in the Monte Carlo search tree, keeps track of the children states"""
-
- def __init__(self, parent=None, state=None, U=0, N=0):
- self.__dict__.update(parent=parent, state=state, U=U, N=N)
- self.children = {}
- self.actions = None
-
-
-def ucb(n, C=1.4):
- return (float('inf') if n.N == 0 else
- n.U / n.N + C * math.sqrt(math.log(n.parent.N) / n.N))
-
-
# ______________________________________________________________________________
# Useful Shorthands
diff --git a/utils4e.py b/utils4e.py
index 792fa9e22..3dfd6c100 100644
--- a/utils4e.py
+++ b/utils4e.py
@@ -1,4 +1,4 @@
-"""Provides some utilities widely used by other modules"""
+"""Provides some utilities widely used by other modules."""
import bisect
import collections
@@ -13,6 +13,8 @@
import numpy as np
+inf = float('inf')
+
# part1. General data structures and their functions
# ______________________________________________________________________________
@@ -22,8 +24,7 @@
class PriorityQueue:
- """A Queue in which the minimum (or maximum) element (as determined by f and
- order) is returned first.
+ """A Queue in which the minimum (or maximum) element (as determined by f and order) is returned first.
If order is 'min', the item with minimum f(x) is
returned first; if order is 'max', then it is the item with maximum f(x).
Also supports dict-like lookup."""
@@ -153,6 +154,13 @@ def powerset(iterable):
return list(chain.from_iterable(combinations(s, r) for r in range(len(s) + 1)))[1:]
+def extend(s, var, val):
+ """Copy dict s and extend it by setting var to val; return copy."""
+ s2 = s.copy()
+ s2[var] = val
+ return s2
+
+
# ______________________________________________________________________________
# argmin and argmax
@@ -201,7 +209,7 @@ def histogram(values, mode=0, bin_function=None):
return sorted(bins.items())
-def dotproduct(X, Y):
+def dot_product(X, Y):
"""Return the sum of the element-wise product of vectors X and Y."""
return sum(x * y for x, y in zip(X, Y))
@@ -231,11 +239,7 @@ def matrix_multiplication(X_M, *Y_M):
def _mat_mult(X_M, Y_M):
"""Return a matrix as a matrix-multiplication of two matrices X_M and Y_M
- >>> matrix_multiplication([[1, 2, 3],
- [2, 3, 4]],
- [[3, 4],
- [1, 2],
- [1, 0]])
+ >>> matrix_multiplication([[1, 2, 3], [2, 3, 4]], [[3, 4], [1, 2], [1, 0]])
[[8, 8],[13, 14]]
"""
assert len(X_M[0]) == len(Y_M)
@@ -607,7 +611,7 @@ def vector_clip(vector, lowest, highest):
# ______________________________________________________________________________
# Misc Functions
-class injection():
+class injection:
"""Dependency injection of temporary values for global functions/classes/etc.
E.g., `with injection(DataBase=MockDataBase): ...`"""
@@ -936,6 +940,21 @@ def __hash__(self):
return 1
+# ______________________________________________________________________________
+# Monte Carlo tree node and ucb function
+class MCT_Node:
+ """Node in the Monte Carlo search tree, keeps track of the children states"""
+
+ def __init__(self, parent=None, state=None, U=0, N=0):
+ self.__dict__.update(parent=parent, state=state, U=U, N=N)
+ self.children = {}
+ self.actions = None
+
+
+def ucb(n, C=1.4):
+ return inf if n.N == 0 else n.U / n.N + C * math.sqrt(math.log(n.parent.N) / n.N)
+
+
# ______________________________________________________________________________
# Useful Shorthands
diff --git a/viterbi_algorithm.ipynb b/viterbi_algorithm.ipynb
new file mode 100644
index 000000000..9c23c4f75
--- /dev/null
+++ b/viterbi_algorithm.ipynb
@@ -0,0 +1,418 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "# Probabilistic Reasoning over Time\n",
+ "---\n",
+ "# Finding the Most Likely Sequence with Viterbi Algorithm\n",
+ "\n",
+ "## Introduction\n",
+ "An ***Hidden Markov Model*** (HMM) network is parameterized by two distributions:\n",
+ "\n",
+ "- the *emission or sensor probabilties* giving the conditional probability of observing evidence values for each hidden state;\n",
+ "- the *transition probabilities* giving the conditional probability of moving between states during the sequence. \n",
+ "\n",
+ "Additionally, an *initial distribution* describes the probability of a sequence starting in each state.\n",
+ "\n",
+ "At each time $t$, $X_t$ represents the *hidden state* and $E_t$ represents an *observation* at that time."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from probability import *"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mclass\u001b[0m \u001b[0mHiddenMarkovModel\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"A Hidden markov model which takes Transition model and Sensor model as inputs\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtransition_model\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msensor_model\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mprior\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransition_model\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtransition_model\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msensor_model\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msensor_model\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprior\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mprior\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;36m0.5\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0.5\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mdef\u001b[0m \u001b[0msensor_dist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mev\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mev\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msensor_model\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msensor_model\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource HiddenMarkovModel"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Finding the Most Likely Sequence\n",
+ "\n",
+ "There is a linear-time algorithm for finding the most likely sequence: the easiest way to think about the problem is to view each sequence as a path through a graph whose nodes are the possible states at each time step. Now consider the task of finding the most likely path through this graph, where the likelihood of any path is the product of the transition probabilities along the path and the probabilities of the given observations at each state. There is a recursive relationship between most likely paths to each state $x_{t+1}$ and most likely paths to each state $x_t$ . We can write this relationship as an equation connecting the probabilities of the paths:\n",
+ "\n",
+ "$$ \n",
+ "\\begin{align*}\n",
+ "m_{1:t+1} &= \\max_{x_{1:t}} \\textbf{P}(\\textbf{x}_{1:t}, \\textbf{X}_{t+1} | \\textbf{e}_{1:t+1}) \\\\\n",
+ "&= \\alpha \\textbf{P}(\\textbf{e}_{t+1} | \\textbf{X}_{t+1}) \\max_{x_t} \\Big(\\textbf{P}\n",
+ "(\\textbf{X}_{t+1} | \\textbf{x}_t) \\max_{x_{1:t-1}} P(\\textbf{x}_{1:t-1}, \\textbf{x}_{t} | \\textbf{e}_{1:t})\\Big)\n",
+ "\\end{align*}\n",
+ "$$\n",
+ "\n",
+ "The *Viterbi algorithm* is a dynamic programming algorithm for *finding the most likely sequence of hidden states*, called the Viterbi path, that results in a sequence of observed events in the context of HMMs.\n",
+ "This algorithms is useful in many applications, including *speech recognition*, where the aim is to find the most likely sequence of words, given a series of sounds and the *reconstruction of bit strings transmitted over a noisy channel*."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "\u001b[0;32mdef\u001b[0m \u001b[0mviterbi\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mHMM\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mev\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m\"\"\"\u001b[0m\n",
+ "\u001b[0;34m [Equation 15.11]\u001b[0m\n",
+ "\u001b[0;34m Viterbi algorithm to find the most likely sequence. Computes the best path and the\u001b[0m\n",
+ "\u001b[0;34m corresponding probabilities, given an HMM model and a sequence of observations.\"\"\"\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mt\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mev\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mev\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mev\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mev\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minsert\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mm\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0.0\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0.0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0m_\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mev\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# the recursion is initialized with m1 = forward(P(X0), e1)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mHMM\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mHMM\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprior\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mev\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# keep track of maximizing predecessors\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbacktracking_graph\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mt\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0melement_wise_product\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mHMM\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msensor_dist\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mev\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m \u001b[0;34m+\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0melement_wise_product\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mHMM\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransition_model\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0melement_wise_product\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mHMM\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransition_model\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mbacktracking_graph\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0melement_wise_product\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mHMM\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransition_model\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0melement_wise_product\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mHMM\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransition_model\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# computed probabilities\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mml_probabilities\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;36m0.0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mev\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# most likely sequence\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mml_path\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mlen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mev\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# the construction of the most likely sequence starts in the final state with the largest probability, and\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;31m# runs backwards; the algorithm needs to store for each xt its predecessor xt-1 maximizing its probability\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mi_max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0margmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mm\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mt\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m-\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mml_probabilities\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mm\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi_max\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mml_path\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mi_max\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m \u001b[0;32melse\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mi\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0mi_max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbacktracking_graph\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi_max\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\n",
+ "\u001b[0;34m\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0mml_path\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mml_probabilities\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n"
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "%psource viterbi"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Umbrella World\n",
+ "---\n",
+ "\n",
+ "> You are the security guard stationed at a secret under-ground installation. Each day, you try to guess whether it’s raining today, but your only access to the outside world occurs each morning when you see the director coming in with, or without, an umbrella.\n",
+ "\n",
+ "In this problem $t$ corresponds to each day of the week, the hidden state $X_t$ represent the *weather* outside at day $t$ (whether it is rainy or sunny) and observations record $E_t$ whether at day $t$ the security guard sees the director carrying an *umbrella* or not."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "#### Observation Emission or Sensor Probabilities $P(E_t := Umbrella_t | X_t := Weather_t)$\n",
+ "We need to assume that we have some prior knowledge about the director's behavior to estimate the emission probabilities for each hidden state:\n",
+ "\n",
+ "| | $yes$ | $no$ |\n",
+ "| --- | --- | --- |\n",
+ "| $Sunny$ | 0.10 | 0.90 |\n",
+ "| $Rainy$ | 0.80 | 0.20 |\n",
+ "\n",
+ "#### Initial Probability $P(X_0 := Weather_0)$\n",
+ "We will assume that we don't know anything useful about the likelihood of a sequence starting in either state. If the sequences start each week on Monday and end each week on Friday (so each week is a new sequence), then this assumption means that it's equally likely that the weather on a Monday may be Rainy or Sunny. We can assign equal probability to each starting state:\n",
+ "\n",
+ "| $Sunny$ | $Rainy$ |\n",
+ "| --- | ---\n",
+ "| 0.5 | 0.5 |\n",
+ "\n",
+ "#### State Transition Probabilities $P(X_{t} := Weather_t | X_{t-1} := Weather_{t-1})$\n",
+ "Finally, we will assume that we can estimate transition probabilities from something like historical weather data for the area. Under this assumption, we get the conditional probability:\n",
+ "\n",
+ "| | $Sunny$ | $Rainy$ |\n",
+ "| --- | --- | --- |\n",
+ "|$Sunny$| 0.70 | 0.30 |\n",
+ "|$Rainy$| 0.30 | 0.70 |"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "umbrella_transition = [[0.7, 0.3], [0.3, 0.7]]\n",
+ "umbrella_sensor = [[0.9, 0.2], [0.1, 0.8]]\n",
+ "umbrellaHMM = HiddenMarkovModel(umbrella_transition, umbrella_sensor)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from graphviz import Digraph"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ "\n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "execution_count": 11,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "dot = Digraph()\n",
+ "\n",
+ "dot.node('I', 'Start', shape='doublecircle')\n",
+ "dot.node('R', 'Rainy')\n",
+ "dot.node('S','Sunny')\n",
+ "\n",
+ "dot.edge('I', 'R', label='0.5')\n",
+ "dot.edge('I', 'S', label='0.5')\n",
+ "\n",
+ "dot.edge('R', 'S', label='0.2')\n",
+ "dot.edge('S', 'R', label='0.4')\n",
+ "\n",
+ "dot.node('Y', 'Yes')\n",
+ "dot.node('N', 'No')\n",
+ "\n",
+ "dot.edge('R', 'R', label='0.6')\n",
+ "dot.edge('R', 'Y', label='0.8')\n",
+ "dot.edge('R', 'N', label='0.2')\n",
+ "\n",
+ "dot.edge('S', 'S', label='0.8')\n",
+ "dot.edge('S', 'Y', label='0.1')\n",
+ "dot.edge('S', 'N', label='0.9')\n",
+ "\n",
+ "dot"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Suppose that $[true, true, false, true, true]$ is the umbrella sequence for the security guard’s first five days on the job. What is the weather sequence most likely to explain this?"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "from utils import rounder"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "([1, 1, 0, 1, 1], [0.8182, 0.5155, 0.1237, 0.0334, 0.021])"
+ ]
+ },
+ "execution_count": 14,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "umbrella_evidence = [True, True, False, True, True]\n",
+ "\n",
+ "rounder(viterbi(umbrellaHMM, umbrella_evidence))"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.7.3"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}