Pacman: Clean implementations of Distributed Constraint Optimization Algorithms
pacman
is available on GitHub at http://github.com/chufansuki/pacman
Main Features
Generation of standard DCOP test problems: weighted graph coloring problems, scale-free problems and related parameter settings to generate the corresponding XML files/YAML files for the problems.
Test file parsing: parse the XML file/YAML file to map the information in the DCOP problem to the underlying structure.
Inter-process communication mechanism: implementation of process communication to simulate multi-intelligent body parallel interaction, and implementation of asynchronous and synchronous control methods.
Multi-platform : Pacman can run on windows, Mac and Linux.
Reproduce several classical DCOP algorithms methods, including MGM, DPOP, etc.
Define algorithm performance metrics and implement each performance metric for Visualization interface for statistical testing of algorithm performance and display of algorithm parameters and performance.