Disclaimer: This repository was created on the 14th January, 2023. It was given as an assignment and I am only sharing my solutions after getting the results. I do not endorse copying or plagiarism.
Score Received: 89/100.
Model requires optimisation.
The goal is to create a reinforcement learning model for Pacman using Markov decision processes (MDP) such as value iteration or policy iteration. The MDP model should work both on medium sized maps and small maps. This assignment prohibits the use of any other learning algorithm and strictly requires the use of value or policy iteration.
You can find the rest of the Pacman game files on UC Berkley's official website. You only need to move the mdpAgents.py file to the pacman game folder, where all the other agents are.
The file is documented, thus, you can easily read through it to understand each step of my solution.