Author: James Long
Date: 13:57:04 08/03/98
As I rewrite my program, I've been seriously considering trying a genetic algorithm to train the evaluation function. I've got a pretty good grasp on how to handle this, but a couple of things are still bothering me: Most programmers hand tune their eval functions, either themselves or with the help of a strong player. Is this because the results are generally better, or because a good learning eval is difficult to write? Dr. Hyatt - do you think Crafty may be in a state of local optima? (Or do you use a learning eval?) Bruce? Anybody? :-) I'm wondering how many programs get caught up on this. For the program to "learn," the eval variables must mutate dependent on the outcome of a game. One approach I read about is to set up a tournament of 100 "players," each player having random values assigned to its eval terms (with the exception of the pawn score). When one player finally wins, the weakest player's terms are modified. This requires thousands of games, but I'm in no hurry. My question is - how to go about modifying the eval terms. I've got a couple ideas, but I'd like to read more. Anybody got some good pointers?? --- James
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.