My second submission to the Pacman coursework in Machine Learning class at King's College London.
This coursework solves the Pacman game using a basic SARSA reinforcement learning algorithm. Details of SARSA: https://en.wikipedia.org/wiki/State%E2%80%93action%E2%80%93reward%E2%80%93state%E2%80%93action
The implementation is in the "mlLearningAgents.py" file.
To run the algorithm, type into the terminal:
python pacman.py -p QLearnAgent -x 2000 -n 2010 -l smallGrid
The algorithm will be trained on 2000 games and will be tested on 10 games. smallGrid is the map.
This coursework is based on the PacmanAI from UC Berkeley: http://ai.berkeley.edu/project_overview.html