Assignments for my first-year MS Intro to AI class at WashU. It is mostly Pacman, which honestly made the algorithms easier to care about..
Small Python warm-up before the Pacman codebase. Mostly lists, loops, dictionaries, recursion, and a tiny fruit-shop problem.
Pacman learns how to move through mazes with DFS, BFS, UCS, and A*. This was the first homework that felt like actual AI.
Ghosts start making choices too, so Pacman needs reflex rules, minimax, alpha-beta, and expectimax.
Gridworld and crawler experiments for value iteration and Q-learning. This is where rewards start turning into behavior.
Invisible ghosts, noisy distances, and belief tracking. Pacman has to hunt using probabilities instead of just seeing everything.
Two agents, one simple team plan: one goes for food, one tries to defend. Very simple strategy, but it runs.
