Behavior-determining AI in Computer Games: A Literature Review

We studied different kinds of algorithm used in games that improve the performance of the AIs, including the frequency-adjusted Q-Learning (FAQ Learning), the robust multi-agent Q-Learning (RoM-Q Learning), and deep recurrent Q-Learning (DRQN). We reviewed a paper for each one of the algorithms and create this review paper presenting the use of different Q-learning algorithms on AI to improve their performance in games. Through improving the performing of AI in these areas, we are able to find something attractive in the different kind of behavior the AI does which is stunning and surprising. So, we can have a more intelligent game systems and become more real.