Competitive Goal-Seeking Through Evolutionary Programming

摘要

The investigation was to examine various ways in which finite-state machines and related mathematical techniques can be used to increase the likelihood of successful game-playing. More specifically, evolutionary programming offers a means for modeling the 'strategy' of the opponent, thus permitting the improvement of play against an intelligent adversary. Mathematical experiments performed in computer simulation were used to demonstrate this capability. Additional experiments concerned pursuit/evasion games similar to those which arise in tactical situations. (Author)