Pacman

Final Project for UMBC CMSC 471 Artificial Intelligence

Michael Kelbaugh, Matthew Landen, Celestine Wong

UC Berkeley Pac-Man Projects

http://ai.berkeley.edu/project_overview.html

All you need to run the code is a standard Python 2.7 installation.

If you get errors about not having the necessary Tkinter module (for graphics),

you can install it from here: http://tkinter.unpythonic.net/wiki/How_to_install_Tkinter

Basic commands to get started:

play the Pacman game manually:

cd Pacman/search python pacman.py

watch the Pacman reflex agent:

cd Pacman/multiagent python pacman.py -p ReflexAgent --frameTime 0.1 -k 1

watch the Pacman minimax agent:

cd Pacman/multiagent python pacman.py -p MinimaxAgent --frameTime 0.1 -k 1

watch Pacman train with reinforcement learning:

cd Pacman/reinforcement python pacman.py -p ApproximateQAgent -a extractor=SimpleExtractor -n 60 -l mediumClassic

watch Pacman play after reinforcement learning:

cd Pacman/reinforcement python pacman.py -p ApproximateQAgent -a extractor=SimpleExtractor -x 50 -n 60 -l mediumClassic