Application of UCT Search to the Connection Games of Hex, Y, *Star, and Renkula!
暂无分享,去创建一个
Play-out analysis has proved a succesful approach for artifi cial intelligence (AI) in many board games. The idea is to play numerous times from the current state to th e end, with randomness in each play-out; a good next move is then chosen by analyzing the set of play-ou ts and their outcomes. In this paper we apply play-out analysis to so-called ‘connection games’ , abstract board games where connectivity of pieces is important. In this class of games, evaluating th e game state is difficult and standard alphabeta search based AI does not work well. Instead, we use UCT se arch, a play-out analysis method where the first moves in the lookahead tree are seen as multi-a rmed bandit problems and the rest of the play-out is played randomly using heuristics. We demons trate the effectiveness of UCT in four different connection games, including a novel game called R nkula!.
[1] Jack van Rijswijck. Search and evaluation in Hex , 2002 .
[2] Peter Auer,et al. Finite-time Analysis of the Multiarmed Bandit Problem , 2002, Machine Learning.
[3] Cameron Browne,et al. Connection Games - Variations on a Theme , 2005 .
[4] Olivier Teytaud,et al. Modification of UCT with Patterns in Monte-Carlo Go , 2006 .
[5] Csaba Szepesvári,et al. Bandit Based Monte-Carlo Planning , 2006, ECML.