UCT Algorithm in Imperfect Information Multi-Player Military Chess Game

DOI

10.2991/jcis.2008.120How to use a DOI?

Keywords

computer game, huge branches, UCT, game tree search, Monte-Carlo, Military Chess

Abstract

UCT (Upper Confidence Bound Apply to Tree) is a new algorithm applies to huge branches mini-max game tree search which gains a great success on a Go program called MoGo. This paper proposes an imperfect information multi-player Military Chess system basing on UCT algorithm which also adopts Monte- Carlo sampling and some other classic game algorithms (alpha-beta pruning, TD (?) Learning and history heuristic). The realization and modification of UCT for Military Chess application are also explained in this paper which improves significantly the performance of the system. Basing on the results of the experiments, the performance of modified UCT for imperfect information game system is discussed.

Copyright

© 2008, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Jiajia Zhang
AU  - Xuan Wang
AU  - Jing Lin
AU  - Zhoayang Xu
PY  - 2008/12
DA  - 2008/12
TI  - UCT Algorithm in Imperfect Information Multi-Player Military Chess Game
BT  - Proceedings of the 11th Joint Conference on Information Sciences (JCIS 2008)
PB  - Atlantis Press
SP  - 719
EP  - 727
SN  - 1951-6851
UR  - https://doi.org/10.2991/jcis.2008.120
DO  - 10.2991/jcis.2008.120
ID  - Zhang2008/12
ER  -