Pacman AI

A Python implementation of artificial intelligence search algorithms to solve problems within the Berkeley Pac-Man environment. The Pac-Man Projects, developed at UC Berkeley, apply AI concepts to the classic arcade game. I help Pac-Man find food, avoid ghosts, and maximise his game score using uninformed and informed state-space search, probabilistic inference, and reinforcement learning.

1. Search Project
   • Implemented DFS,
   • BFS,
   • UCS,
   • Greedy Search,
   • A* Search
2. MultiAgent Project (Adversarial search)
   • Minimax
   • Alpha-beta pruning
   • Expectimax
3. Markov Decision Processes & Reinforcemenet Learning
   • Value Iteration
   • Policy Iteration
   • Asynchronous value iteration
   • Prioritized sweeping value iteration
   • Epsilon greedy
   • Q-learning
   • Approximate Q-learning