AI Final Project. Pacman tournament that pits the following different kinds of agents against each other: -Baseline agents -Greedy agents -Q-learning agents -Inference agents

We found inference agents strongly outperformed other kinds of agents.

The basic command to run the code is the following:

python capture.py

to specify red and blue teams, add either "-r" followed by the agent or "-b", followed by the agent.

For example to run a red inference team against a blue greedy team:

python capture.py -r inferenceTeam -b greedyTeam

The team names for each agent we implemented are: inferenceTeam -> inferenceTeam.py greedyTeam -> greedyTeam.py qlearnTeam -> qlearnTeam.py

The provided "Base line team" can be ran by "baselineTeam"

Thanks to UC Berkeley for the API from here https://inst.eecs.berkeley.edu/~cs188/sp09/projects/contest/contest.html.