基於UCT之九路電腦圍棋程式HappyGO的設計與實作

Design and Implementation of 9x9 Computer Go Program HappyGO Base on UCT

王永樂(Yung-Le Wang)

指導教授 ： 吳毅成

國立交通大學/資訊學院/資訊學院碩士在職專班資訊組/碩士(2009年)

https://doi.org/10.6842/NCTU.2009.00104

摘要

UCT演算法運用在電腦圍棋程式上只需要基本圍棋知識的特性，提供有別於傳統圍棋程式的另一種程式設計與實作的方法。使用UCT演算法實作圍棋程式已成目前最主要的方法之一。在本論文中，將說明如何基於UCT演算法實作九路圍棋程式HappyGO並加入不同的圍棋策略。

關鍵字

圍棋 ； 九路圍棋程式 ； 蒙地卡羅

並列摘要

Only basic Go knowledge is needed to implement computer Go program based on UCT algorithm. This feature provides another method that is different to traditional Go program to design and implement Go program. Using UCT algorithm is one of major methods to implement Go program. In this paper, we give detailed information about how to implement a Go program, HappyGO, based on the UCT algorithm added with some Go strategies.

並列關鍵字

GO ； UCT ； HappyGO

參考文獻

1. D.J. H. Brown, Dowsey, S. “The challenge of Go.”New Scientist,

5. N.Metropolis, S.Ulam, “The Monte Carlo Method”,Journal of the American Statistical Association, Vol.44, No.247, pp.335-341, 1949

12. Gelly,S.,Silver,D.,”Combining online and offline learning in UCT”,17th International Conference on Machine Learning, pp.273-280, 2007

Vol.81, pp.303-305,1979

2. Zobrist, A. L. “Feature Extraction and Representation for Pattern