Reinforcement-learning-for-Pacman

Q learning with function approximation

Read the document, called Document.pdf, which explains how it works.

My Q-learning is in mlLearningAgents.py

Work in python 2.

Execute in terminal

Play 2010 games and use the first 2000 games to train in a smallGrid.

python pacman.py -p QLearnAgent -x 2000 -n 2010 -l smallGrid

Play 110 games and use the first 100 games to train in a mediumGrid.

python pacman.py -p QLearnAgent -x 100 -n 110