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. ...

• Type: Conference
• Year: 2004
• Where: IWEC
• Authors: Sander Bakkes, Pieter Spronck, Eric O. Postma
• Keywords: Adaptive Mechanism, Computer Games, Computer-controlled Opponents, IWEC 2004, IWEC 2007