AI-homework

This repository contains four separate projects written in AI class.

The first one is search. In this project, I use deep-first search, breadth-first search, uniform-cost search and A* search to implement a Pacman AI agent.

The second one is CSP(Constraint satisfaction problem). I use Backtracking search and arc consistency algorithm(AC3) to solve N-queen problem.

The third project is about MDP(Markoff decision process) and Reinforcement Learning. I write an RL algorithm (feature-based, model free Q-Learning) to implement an agent for the poker game: Black-Jack.

The forth project is about inference. In this project, I implemented an HMM model to infer the probability of other cars and obstacles. This will help the car to auto-drive, and implemented a particle-filtering algorithm to reduce computing cost.