Evolutionary algorithms for board game players with domain knowledge

Incorporating a priori knowledge, such as expert knowledge, metaheuristics, human preferences, and most importantly domain knowledge discovered during evolutionary search, into evolutionary algorithms has gained increasing interest in recent years. In this chapter, we present a method for systematically inserting expert knowledge into evolutionary board game framework at the opening, middle, and endgame stages. In the opening stage, openings defined by the experts are used. In this work, we use speciation techniques to search for diverse strategies that embody different styles of game play and combine them using voting for higher performance. This idea comes from the common knowledge that the combination of diverse well-playing strategies can defeat the best one because they can complement each other for higher performance. Finally, we use an endgame database. Experimental results on checkers and Othello games show that the proposed method is promising to evolve better strategies.