MA-FD code, with the option of using selfish agents.
Instructions on running an example problem. go to the src/ directory ./build_all ./translate/translate.py ../benchmarks/rovers/domain.pddl ../benchmarks/rovers/p12.pddl./preprocess/preprocess < output.sas This specific problem has 4 agents called "rover[0-3]", so open agents file and insert the following: 4 rover0 rover1 rover2 rover3
The names are important, they are how the actions are distributed among the agents. Then go to comm file and insert the following: 127.0.0.1:3010 127.0.0.1:3011 127.0.0.1:3012 127.0.0.1:3013
This gives an ip address to each of the 4 agents
Now run the 4_agents script, which contains the following text: ./search/downward --search "astar(lmcut())" --agents 0 < output > 0 & ./search/downward --search "astar(lmcut())" --agents 1 < output > 1 & ./search/downward --search "astar(lmcut())" --agents 2 < output > 2 & ./search/downward --search "astar(lmcut())" --agents 3 < output > 3 python parse_results.py 4
See other scripts on ways of running non-optimal algorithm and other heuristics.