Fuzzeval: A Fuzzy Controller-Based Approach in Adaptive Learning for Backgammon Game

In this paper we investigate the effectiveness of applying fuzzy controllers to create strong computer player programs in the domain of backgammon. Fuzzeval, our proposed mechanism, consists of a fuzzy controller that dynamically evaluates the perceived strength of the board configurations it re-ceives. Fuzzeval employs an evaluation function that adjusts the membership functions linked to the linguistic variables in the knowledge base. The mem-bership functions are aligned to the average crisp input that was successfully used in the past winning games. Fuzzeval mechanisms are adaptive and have the simplicity associated with fuzzy controllers. Our experiments show that Fuzzeval improves its performance up to 42% after a match of only one hun-dred backgammon games played against Pubeval, a strong intermediate level program.