第13卷第4期 智能系统学报 Vol.13 No.4 2018年8月 CAAI Transactions on Intelligent Systems Aug.2018 D0:10.11992/tis.201609023 网络出版地址：http:/kns.cnki.net/kcms/detail/23.1538.TP.20180515.1116.002.html 基于棋型的藏族“久”棋计算机博弈研究 李霞丽，吴立成，李永集 (中央民族大学信息工程学院，北京100081) 摘要：“久”棋是藏族人民的传统棋类游戏，游戏过程分为布局阶段和战斗阶段，布局的质量对弈棋结果影响 很大。与围棋博弈智能软件战胜人类高手的情况比较，“久”棋博弈研究几乎空白。为了拓宽机器博弈研究的 游戏范围，开发具有较高棋力的“久”棋软件，作者开展了基于棋型的“久”棋计算机博弈研究。通过实地考察， 在四川阿坝地区采集了约300局有效的“久”棋对弈数据，提取了常见棋型，分别为棋型命名为三角、三子、二 子、对角、四子等。在布局阶段，采用模式匹配算法提高棋型的匹配速度。在布局和战斗阶段，基于棋型，设计 了具有优先级别的防守、攻击、连子策略。采用C语言开发了“久”棋博弈软件，该软件具有人人对弈、人机对 弈、自动录制棋谱等功能。该软件在2016年四川省阿坝县第七届“体彩杯”藏棋比赛中成功开展了人机对弈， 但是棋力有待提高。结果表明，基于棋型的攻防策略能够有效地应用于“久”棋计算机博弈。 关键词：计算机博弈：藏族“久”棋：棋型：攻防策略：模式匹配 中图分类号：TP39文献标志码：A文章编号：1673-4785(2018)04-0577-07 中文引用格式：李霞丽，吴立成，李永集.基于棋型的藏族“久”棋计算机博弈研究J智能系统学报，2018,13(4)：577-583， 英文引用格式：LI Xiali,WU Licheng,LI Yongji.Tibetan JIU computer game research based on chess formJ.CAAI transactions on intelligent systems,2018,13(4):577-583. Tibetan JIU computer game research based on chess form LI Xiali,WU Licheng,LI Yongji (School of Information Engineering,Minzu University of China,Beijing 100081,China) Abstract:JIU is a traditional Tibetan board game that is divided into two sequential stages-embattle(or prepare for battle)and battle.The embattle stage has a critical effect on the subsequent battle.Compared with AlphaGo and Al- phaGo Zero computer programs,which have defeated top human players,research on the game of JIU is almost nonex- istent.To broaden the scope of computer game research and work toward the development of a sophisticated JIU chess game,we conducted a computer game study of the formations used in chess.Specifically,we collected about 300 JIU play records in an on-the-spot investigation in the Aba Autonomous Prefecture of Sichuan Province.In our analysis of these play records,we identified several common chess formations,which we refer to as the triangle,trinity,twain,con- trast,and square formations.To increase the speed of the character string matching process,we used a pattern matching algorithm in the embattle stage.We also designed defensive,attack,and collaboration strategies for the embattle and battle stages based on these chess formations.The defensive,attack,and collaboration strategies have decreasing prior- ity.Then,we developed JIU chess software using C language,with functions including the man-man VS mode, human-computer VS AI mode,and automatic recording of the play process.This software performed consistently in the man-machine game play exhibition at the 2016 Seventh "Sports Lottery Cup"Tibetan Chess Contest held in the Aba Autonomous Prefecture of Sichuan Province.However,the chess level realized by the software must be improved.The results show that attack and defense strategies based on chess formations can be effectively applied to JIU chess com- puter games. Keywords:computer games,Tibetan JIU chess;chess form;attack and defense strategies,pattern matching 收稿日期：2016-09-23.网络出版日期：2018-05-17 藏棋是藏族人民的传统游戏，主要流传于我 基金项目：国家自然科学基金项目(61602539,61773416) 通信作者：吴立成.E-mail:Wulicheng@tsinghua.edu.cn 国西藏以及十大藏区，是藏族文化宝库中一颗璀DOI: 10.11992/tis.201609023 网络出版地址: http://kns.cnki.net/kcms/detail/23.1538.TP.20180515.1116.002.html 基于棋型的藏族“久”棋计算机博弈研究 李霞丽，吴立成，李永集 （中央民族大学 信息工程学院，北京 100081） 摘 要：“久”棋是藏族人民的传统棋类游戏，游戏过程分为布局阶段和战斗阶段，布局的质量对弈棋结果影响 很大。与围棋博弈智能软件战胜人类高手的情况比较，“久”棋博弈研究几乎空白。为了拓宽机器博弈研究的 游戏范围，开发具有较高棋力的“久”棋软件，作者开展了基于棋型的“久”棋计算机博弈研究。通过实地考察， 在四川阿坝地区采集了约 300 局有效的“久”棋对弈数据，提取了常见棋型，分别为棋型命名为三角、三子、二 子、对角、四子等。在布局阶段，采用模式匹配算法提高棋型的匹配速度。在布局和战斗阶段，基于棋型，设计 了具有优先级别的防守、攻击、连子策略。采用 C 语言开发了“久”棋博弈软件，该软件具有人人对弈、人机对 弈、自动录制棋谱等功能。该软件在 2016 年四川省阿坝县第七届“体彩杯”藏棋比赛中成功开展了人机对弈， 但是棋力有待提高。结果表明，基于棋型的攻防策略能够有效地应用于“久”棋计算机博弈。 关键词：计算机博弈；藏族“久”棋；棋型；攻防策略；模式匹配 中图分类号：TP39 文献标志码：A 文章编号：1673−4785(2018)04−0577−07 中文引用格式：李霞丽, 吴立成, 李永集. 基于棋型的藏族“久”棋计算机博弈研究[J]. 智能系统学报, 2018, 13(4): 577–583. 英文引用格式：LI Xiali, WU Licheng, LI Yongji. Tibetan JIU computer game research based on chess form[J]. CAAI transactions on intelligent systems, 2018, 13(4): 577–583. Tibetan JIU computer game research based on chess form LI Xiali，WU Licheng，LI Yongji (School of Information Engineering, Minzu University of China, Beijing 100081, China) Abstract: JIU is a traditional Tibetan board game that is divided into two sequential stages—embattle (or prepare for battle) and battle. The embattle stage has a critical effect on the subsequent battle. Compared with AlphaGo and Al￾phaGo Zero computer programs, which have defeated top human players, research on the game of JIU is almost nonex￾istent. To broaden the scope of computer game research and work toward the development of a sophisticated JIU chess game, we conducted a computer game study of the formations used in chess. Specifically, we collected about 300 JIU play records in an on-the-spot investigation in the Aba Autonomous Prefecture of Sichuan Province. In our analysis of these play records, we identified several common chess formations, which we refer to as the triangle, trinity, twain, con￾trast, and square formations. To increase the speed of the character string matching process, we used a pattern matching algorithm in the embattle stage. We also designed defensive, attack, and collaboration strategies for the embattle and battle stages based on these chess formations. The defensive, attack, and collaboration strategies have decreasing prior￾ity. Then, we developed JIU chess software using C language, with functions including the man–man VS mode, human–computer VS AI mode, and automatic recording of the play process. This software performed consistently in the man–machine game play exhibition at the 2016 Seventh “Sports Lottery Cup” Tibetan Chess Contest held in the Aba Autonomous Prefecture of Sichuan Province. However, the chess level realized by the software must be improved. The results show that attack and defense strategies based on chess formations can be effectively applied to JIU chess com￾puter games. Keywords: computer games; Tibetan JIU chess; chess form; attack and defense strategies; pattern matching 藏棋是藏族人民的传统游戏，主要流传于我 国西藏以及十大藏区，是藏族文化宝库中一颗璀 收稿日期：2016−09−23. 网络出版日期：2018−05−17. 基金项目：国家自然科学基金项目 (61602539, 61773416). 通信作者：吴立成. E-mail：Wulicheng@tsinghua.edu.cn. 第 13 卷第 4 期 智 能 系 统 学 报 Vol.13 No.4 2018 年 8 月 CAAI Transactions on Intelligent Systems Aug. 2018