OPTIMIZING GoTools' SEARCH HEURISTICS USING GENETIC ALGORITHMS

Article type: Review Article

Authors: Pratola, Matthew^{; 1} | Wolf, Thomas^{; 2}

Affiliations: St. Catharines, Ontario, Canada

Note: [1] Department of Computer Science, Brock University, Canada, Email: [email protected].

Note: [2] Department of Mathematics, Brock University, Canada, Email: [email protected].

Abstract: GoTools is a program which solves life-and-death problems in the game of Go. This paper describes experiments using a Genetic Algorithm to optimize heuristic weights used by GoTools' tree-search. The complete set of heuristic weights is composed of different subgroups, each of which can be optimized with a suitable fitness function. As a useful side product, an MPI interface for FreePascal (Message Passing Interface) was implemented to allow the use of a parallelized fitness function running on a Beowulf cluster. The aim of this exercise is to optimize the current version of GoTools, and to make tools available in preparation of an extension of GoTools for solving open boundary life-and-death problems, which will introduce more heuristic parameters to be fine-tuned.

DOI: 10.3233/ICG-2003-26106

Journal: ICGA Journal, vol. 26, no. 1, pp. 28-48, 2003