Learning to play Tic-Tac-Toe

Implement an algorithm similar to that discussed for the checkers problem in class, but use the simpler problem of Tic-Tac-Toe. Represent the learned function $V-hat$ as a linear combination of board features of your choice. You may use the LMS weight update algorithm, some other hill-climber of your choice, a genetic algorithm, or some other algorithm of your design. To train your program, play it repeatedly against

• Undergrads: a second copy of your program that uses a fixed evaluation function that you create by hand (static evaluator). This second copy of your program should not lose to me.

• Grads: a second copy of your program.

Good Luck

Turning it in

1. Your source and a script of a sample game of your learned TTT player
2. The plot of the percent of games won by your system versus the number of training games played.
3. If you use an algorithm of your own design, also turn in a description of the algorithm and an analyis of its behavior.