Merge branch 'master' of github.com:rsarai/swarm-algorithms

Author: rsarai

.gitignore

@@ -0,0 +1,129 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/

fish_school_search.py

@@ -1,4 +1,5 @@
 import copy
+import numpy as np
 import functools
 import random
 import math
@@ -7,120 +8,108 @@
 from particle import Fish
 from parameters import num_of_individuos, dimensions, iterations_number
 
+# np.random.seed(42)
 
 class FSS():
 
-    def __init__(self, function_wrapper):
-        self.function = function_wrapper
+    def __init__(self, objective_function):
+        self.function = objective_function
+        self.dimensions = dimensions
+        self.iterations_number = iterations_number
+        self.num_of_individuos = num_of_individuos
+        self.cluster = []
+        self.global_best = float('inf')
+        self.global_best_position = []
+
+        # Params
+        self.total_weight = 1 * self.num_of_individuos
+        self.initial_step_ind = 0.1
+        self.final_step_ind = 0.0001
+        self.step_ind = self.initial_step_ind * (objective_function.upper_bound - objective_function.lower_bound)
+        self.initial_step_vol = 0.01
+        self.final_step_vol = 0.001
+        self.step_vol = self.initial_step_vol * (objective_function.upper_bound - objective_function.lower_bound)
+        self.list_global_best_values = []
 
     def search(self):
-        bests = []
-        maximum_weight_value = 1000
-        individual_step_dacay = 0.00005
-        self._initialize_particles()
-
-        for iterations in range(iterations_number):
-
-            # individual_movement
-            fitness_variations = []
-            for fish in self.population:
-                new_position = fish.individual_movement()
-                new_fish = Fish(fish.function_wrapper, new_position, fish.weights)
-                new_fish.aply_function_on_current_position()
-                fish.aply_function_on_current_position()
-
-                if new_fish.fitness < fish.fitness:
-                    new_fish.previous_position = fish.current_position
-                    fish = new_fish
-                else:
-                    fish.previous_position = fish.current_position
-
-                fitness_variations.append(abs(new_fish.fitness - fish.fitness))
-
-            # feed fishes
-            max_variation = max(fitness_variations)
-            for fish in self.population:
-                new_weight = fish.weights + (fish.fitness / max_variation)
-                fish.previous_weight = fish.weights
-                fish.weights = new_weight
-                if fish.weights > maximum_weight_value:
-                    fish.weights = maximum_weight_value
-
-            # calculates vector of instintict moviments
-            distance_variations_sum = [0 for i in range(0, dimensions)]
-            fitness_variations_sum = sum(fitness_variations)
-            for fish in self.population:
-
-                for i in range(0, dimensions):
-                    distance_variations = abs(fish.current_position[i] - fish.previous_position[i])
-                    distance_variations_sum[i] = distance_variations_sum[i] + (distance_variations * fitness_variations[i])
-
-            instintive_movement_array = []
-            for i in range(0, dimensions):
-                if fitness_variations_sum != 0:
-                    instintive_movement_array.append(distance_variations_sum[i]/fitness_variations_sum)
-                else:
-                    instintive_movement_array.append(0)
-
-            # executes instintive moviments
-            for fish in self.population:
-                for i in range(0, dimensions):
-                    fish.current_position[i] = fish.current_position[i] + instintive_movement_array[i]
-
-            # calculates the baricenter
-            weights_sum = 0
-            for fish in  self.population:
-                weights_sum += fish.weights
-
-            weights_sum_times_position = 0
-            for fish in self.population:
-                for i in range(0, dimensions):
-                    weights_sum_times_position = weights_sum_times_position + (fish.current_position[i] * fish.weights)
-
-            baricenter = [weights_sum_times_position / weights_sum for i in range(0, dimensions)]
-
-            # executes volitive moviment
-            for fish in self.population:
-                distance_to_baricenter = np.linalg.norm([fish.current_position, baricenter])
-
-                volational_array = []
-                difference_to_baricenter = []
-                weight_variation_sum = 0
-                for i in range(0, dimensions):
-                    volational_array.append(2 * fish.current_position[i])
-                    difference_to_baricenter.append(fish.current_position[i] - baricenter[i])
-                    weight_variation_sum = weight_variation_sum + abs(fish.weights - fish.previous_weight)
-
-                sign = functools.partial(math.copysign, 1)
-                for i in range(0, dimensions):
-                    fish.current_position[i] = fish.current_position[i] + -sign(weight_variation_sum) * volational_array[i] * random.random() * difference_to_baricenter[i] / distance_to_baricenter
-
-            for fish in self.population:
-                for i in range(0, dimensions):
-                    fish.current_position[i] -= individual_step_dacay
-
-            for fish in self.population:
-                fish.aply_function_on_current_position()
-
-            bests.append(min([fish.fitness for fish in self.population]))
-        return bests
-
-
-    def _initialize_particles(self):
-        self.population = []
-        for i in range(num_of_individuos):
-            random_position = [self._get_random_number() for index in range(dimensions)]
-            fish = Fish(self.function, random_position, random.uniform(300, 600))
-            self.population.append(fish)
-
-    def _euclidian_distance(self, particle_i, particle_j):
-        total = 0
-        for i in range(dimensions):
-            distance = (particle_i[i] - particle_j[i])**2
-            total += distance
-        return math.sqrt(total)
+        self._initialize_cluster()
+
+        for i in range(self.iterations_number):
+            self.evaluate_cluster()
+            self.updates_optimal_solution()
+
+            self.apply_individual_movement()
+            self.evaluate_cluster()
+            self.updates_optimal_solution()
+
+            self.apply_feeding()
+
+            self.apply_instintive_collective_movement()
+            self.apply_collective_volitive_movement()
+
+            self.update_step(i)
+            self.update_total_weight()
+
+            self.evaluate_cluster()
+            self.updates_optimal_solution()
+            self.list_global_best_values.append(self.global_best)
+            print("iter: {} = cost: {}".format(i, self.global_best))
+
+    def update_total_weight(self):
+        self.total_weight = sum([fish.weight for fish in self.cluster])
+
+    def _initialize_cluster(self):
+        self.cluster = []
+        for _ in range(self.num_of_individuos):
+            fish = Fish(
+                positions=[self._get_random_number() for _ in range(dimensions)],
+                objective_function=self.function
+            )
+            self.cluster.append(fish)
+
+    def evaluate_cluster(self):
+        for fish in self.cluster:
+            fish.evaluate()
+
+    def updates_optimal_solution(self):
+        for fish in self.cluster:
+            if fish.fitness < self.global_best:
+                self.global_best = fish.fitness
+                self.global_best_position = list(fish.current_position)
+
+    def apply_individual_movement(self):
+        for fish in self.cluster:
+            fish.update_position_individual_movement(self.step_ind)
+
+    def apply_feeding(self):
+        max_delta_fitness = max([fish.delta_fitness for fish in self.cluster])
+        for fish in self.cluster:
+            fish.feed(max_delta_fitness)
+
+    def apply_instintive_collective_movement(self):
+        sum_delta_fitness = sum([fish.delta_fitness for fish in self.cluster])
+
+        for fish in self.cluster:
+            fish.update_position_collective_movement(sum_delta_fitness)
+
+    def _calculate_barycenter(self):
+        sum_weights = sum([fish.weight for fish in self.cluster])
+        sum_position_and_weights = [[x * fish.weight for x in fish.current_position] for fish in self.cluster]
+        sum_position_and_weights = np.sum(sum_position_and_weights, 0)
+        return [s / sum_weights for s in sum_position_and_weights]
+
+    def apply_collective_volitive_movement(self):
+        barycenter = self._calculate_barycenter()
+        current_total_weight = sum([fish.weight for fish in self.cluster])
+        search_operator = -1 if current_total_weight > self.total_weight else 1
+        for fish in self.cluster:
+            fish.update_position_volitive_movement(barycenter, self.step_vol, search_operator)
+
+    def update_step(self, current_i):
+        self.step_ind = self.initial_step_ind - current_i * float(
+            self.initial_step_ind - self.final_step_ind) / iterations_number
+        self.step_vol = self.initial_step_vol - current_i * float(
+            self.initial_step_vol - self.final_step_vol) / iterations_number
 
     def _get_random_number(self):
-        return (
-            self.function.lower_bound + random.uniform(0, 1) * (self.function.upper_bound - self.function.lower_bound)
-        )
+        return np.random.uniform(self.function.lower_bound, self.function.upper_bound)