TEAM: The team-oriented evolutionary adaptability mechanism

Many commercial computer games allow a team of players to match their skills against another team, controlled by humans or by the computer. Most players prefer human opponents, since the artificial intelligence of a computer-controlled team is in general inferior. An adaptive mechanism for team-oriented artificial intelligence would allow computer-controlled opponents to adapt to human player behaviour, thereby providing a means of dealing with weaknesses in the game AI. Current commercial computer games lack challenging adaptive mechanisms. This paper proposes "TEAM", a novel team-oriented adaptive mechanism which is inspired by evolutionary algorithms. The performance of TEAM is evaluated in an experiment involving an actual commercial computer game (the Capture The Flag team-based game mode of the popular commercial computer game Quake III). The experimental results indicate that TEAM succeeds in endowing computer-controlled opponents with successful adaptive performance. We therefore conclude that TEAM can be successfully applied to generate challenging adaptive opponent behaviour in team-oriented commercial computer games.