Familiarising Probabilistic Distance Clustering System of Evolving Awale Player

This study developed a new technique based on Probabilistic Distance Clustering (PDC) for evolvingAwale player and to compare its performance with that of a technique based on approximation of minimumand maximum operators with generalized mean-value operator. The basic theory of pd-clustering is basedon the assumption that the probability of an Euclidean point belonging to a cluster is inverselyproportional to its distance from the cluster centroid. Treating game strategies as a vector space model, itis possible to extend pd-clustering technique to game playing by estimating the probability that a givenstrategy is in a certain cluster of game strategies. As a result, the strategy that has the highest probabilityand shortest distance to a cluster of alternative strategies is recommended for the player.