——Project Description——

In this project, our team developed two agents with Q-learning and Deep Q Network (DQN). Each agent has three different behaviors in the game: normal, conservative, and aggressive. Each agent supports playing with human or our evaluation benchmark (see final report for more information).