Course practical for AE4422-20 Agent-based Modelling and Simulation for Air Transportation.
This git repo contains implementations of several path planners that are used to solver an arbitrary multiagent path finding problem. The solutions include:
- a prioritize A* path planner,
- a CBS and CBS cycle limited path planner,
- and a distributed path planner.
To run the code, clone this repository onto your local machine. Open command prompt at the directory where you cloned the repo, and run the command:
python run_experiments.py --solver solver_of_choice --instance instance_of_choice --hvals heuristic_of_choice(optional)
For solver_of_choice you can enter any of the following solvers:
Independent(picks the independent solver, has no collision avoidance)Prioritized(picks the Prioritized A* solver)CBS(picks the conflict based search solver)CBSCL(picks the conflict based search with cycle limiting solver)Distributed(picks the distributed planning solver)
For instance_of_choice you can enter the path to the text file containing the map of your choice, a large collection of maps are presente already in the \instances\ directory. An example of instance_of_choice is:
instances\map1.txt
For heuristics_of_choice, note this is an optional argument, but you can pick to enter either:
old(picks the old heuristics map)goals(picks the heursitics map with extra costs on goal cells)
Example:
python run_experiments.py --solver Distributed --instance instances\map1.txt
Below is a gif demonstrating the CBS cycle limited solver solving a sliding puzzle, with each puzzle piece being an agent.
