Pac-Man AI
Overview
Developed intelligent Pac-Man agents using reinforcement learning and A* algorithms for optimal pathfinding and ghost avoidance across multiple AI strategies.
Deep Dive
This project explored the practical applications of foundational artificial intelligence concepts by building autonomous agents capable of playing Pac-Man at a superhuman level. Starting with classic search algorithms like Depth-First Search (DFS) and Breadth-First Search (BFS), I progressively implemented more sophisticated techniques including A* search with custom heuristics to optimize pathfinding toward food while safely navigating complex mazes. The pinnacle of the project involved designing a Reinforcement Learning agent based on Q-learning. By assigning targeted rewards and penalties, the agent autonomously learned complex strategies - such as safely trailing ghosts or clustering pellets - demonstrating the power of machine learning in dynamic, adverse environments.