Overview

This repository contains the prototype implementation of the proposed Multi-perspective Online Conformance Checking (MOCC) approach presented in the paper An Alignment-based Multi-Perspective Online Conformance Checking Technique by Zsuzsanna Nagy and Agnes Werner-Stark. The paper is available here.

This prototype was implemented using version 2.7.12 of the PM4Py library and the implementation of the proposed incremental A* approach. The source code of the incremental A* algorithm was modified to be able to calculate multi-perspective (prefix-)alignemnts and integrated into this version of PM4Py. Additionally, program codes were developed to allow importing DPN process models from PNML files and to generate and solve a MILP problem for determining the OVAs based on the given variable writings and guard functions.

Repository structure

• pm4py: Modified version of the PM4Py library used in the experiments.
• input: Contains DPN process models and event logs utilized during the experiments.
• experiments.py: Contains the code for implementing and running the experiments.

Changes made to the PM4Py library

• Added files:
  • incremental_a_star_mp.py: Modified version of the incremental A* algorithm used by the MOCC approach to calculate multi-perspective (prefix-)alignments.
  • mp_utils.py: Contains functions to generate and solve a MILP problem for finding the Optimal Variable Assignment (OVA) for an alignment, given a variable writing sequence.
  • incremental_a_star.py: Modified version of the incremental A* algorithm to support both offline and online alignments.
• Modified files:
  • obj.py: Added the DataPetriNet class to handle DPNs (Data Petri Nets).
  • pnml.py: Extended to import DPN process models with initial variable values.
  • petri_utils.py: Extended to support DPNs.
  • align_utils.py: Updated the standard cost function to assign a cost of 1 for every deviation from the model.