Genetic programming applied to strategies learning.

This study addresses the problem of knowledge acquisition to decision taken, applied to Tic-tac-toe game: a fix structure and rules, with a reasonable number of solutions, each one carrying to a different result (win, lost or tie). The comparative study focus the case where all possible states are modelled by genetic programming, and a set of rules are applied against the opponent. The other approach is the acquisition of knowledge on fly, applying each solution for a number of trials, and using genetic programming to obtain the better solution.