Reinforcement Learning SARSA

For more information on project, read Project_report.pdf.

PacMan Game developed using Python.
Implemented SARSA and compared it with Linear Function Approximation.
Coordinated implementation of Alpha Beta Pruning, Min-Max Reflex Agents, Markov Decision Process, Q-Learning, Joint-Particle Filtering to develop a PacMan Game in Python.
Led testing of SARSA algorithm in Crawler with 2 degree of freedom and achieved 7% improvement over Q-learning.

For Grader:
Run AutoGrader.py

For Others:

1. Run Crawler.py to see difference between original and new algorithm.
2. Run Pacman