Developing an Artificial Intelligence Engine

As computer games become more complex and consumers demand more sophisticated computer controlled agents, developers are required to place a greater emphasis on the artificial intelligence aspects of their games. One source of sophisticated AI techniques is the artificial intelligence research community. This paper discusses recent efforts by our group at the University of Michigan Artificial Intelligence Lab to apply state of the art artificial intelligence techniques to computer games. Our experience developing intelligent air combat agents for DARPA training exercises, described in John Laird's lecture at the 1998 Computer Game Developer's Conference, suggested that many principles and techniques from the research community are applicable to games. A more recent project, called the Soar/Games project, has followed up on this by developing agents for computer games, including Quake II and Descent 3. The result of these two research efforts is a partially implemented design of an artificial intelligence engine for games based on well established AI systems and techniques.