Computational Intelligence in Games : The General Video Game AI Competition

A growing section of Artificial Intelligence is Computational Intelligence in Games. Artificial Intelligence methods are applied to computer games in several different ways. In this paper we deal with the General Video Game Playing of computer games where we cannot be sure how the agent is able to score or to win the game at all. Three different approaches that belong to the research areas Heuristic, Reinforcement Learning and Nature Inspired are being considered. A General Video Game Language environment is used to implement our controllers and to simulate a search of possible next game states. The theories behind the different approaches and the modifications to fit to our problem are described. Finally we evaluate the algorithms by using 20 general video games as a test set and compare the average and standard deviation of the winning rate, score and elapsed time.

[1]  Raúl Rojas,et al.  Neural Networks - A Systematic Introduction , 1996 .

[2]  Simon M. Lucas,et al.  A Survey of Monte Carlo Tree Search Methods , 2012, IEEE Transactions on Computational Intelligence and AI in Games.

[3]  Tom Schaul,et al.  A video game description language for model-based or interactive learning , 2013, 2013 IEEE Conference on Computational Inteligence in Games (CIG).

[4]  Risto Miikkulainen,et al.  General Video Game Playing , 2013, Artificial and Computational Intelligence in Games.

[5]  Julian Togelius,et al.  A Panorama of Artificial and Computational Intelligence in Games , 2015, IEEE Transactions on Computational Intelligence and AI in Games.

[6]  Frank Klawonn,et al.  Computational Intelligence: A Methodological Introduction , 2015, Texts in Computer Science.