Improving Opponent Intelligence by Machine Learning

Artificially intelligent opponents in virtual world computer games are almost exclusively controlled by manually-designed scripts. With increasing game complexity, the scripts tend to become quite complex too. As a consequence they often contain “holes” that can be exploited by the human player. The research question addressed in this paper reads: How can machine learning be used to improve the quality of opponent intelligence in computer games? We study the off-line application of evolutionary learning to generate neural-network controlled opponents for a complex strategy game called PICOVERSE. The results show that the evolved opponents outperform a manuallyscripted opponent. In addition, it is shown that evolved opponents are capable of identifying and exploiting holes in a scripted opponent. We conclude that machine learning is potentially an effective tool to improve quality of opponent intelligence in computer games.