Abstract
The game of Amazons is a combinatorial game sharing some properties of both chess and Go. We study programs which play Amazons with strategies based on Monte-Carlo Tree Search and a classical search algorithm, Alpha-Beta pruning. We execute several experiments to investigate the effect of increasing the number of searches in a Monte-Carlo Tree Search program. We show that increasing the number of searches is not an efficient method to strengthen the program for Amazons. On the other hand, augmenting the algorithms with a choice of several evaluation functions fulfills has great influence on playing strength.
Author supplied keywords
Cite
CITATION STYLE
Kato, H., Fazekas, S. Z., Takaya, M., & Yamamura, A. (2015). Comparative study of Monte-Carlo tree search and Alpha-Beta pruning in Amazons. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9357, pp. 139–148). Springer Verlag. https://doi.org/10.1007/978-3-319-24315-3_14
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
