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