热门搜索: 论文 发表 社科期刊 北大核心 南大核心 cssci 科技期刊 教育

当前位置:主页 > 科技论文 > 人工智能论文 >

六子棋计算机博弈及其系统的研究与实现

发布时间:2018-10-19 08:46  文章来源:笔耕文化传播
【摘要】: 计算机博弈是人工智能领域一个极其重要且最具挑战性的研究方向之一,它的研究为人工智能带来了很多重要的方法和理论,产生了广泛的社会影响和学术影响以及大量的研究成果。在过去的半个世纪里,世界各地的学者花费了大量的心血对于计算机博弈包括奥赛罗、checker、国际象棋、中国象棋、五子棋、围棋进行研究。这是因为计算机博弈是人工智能的一块试金石,然而棋类游戏又是计算机博弈的一个标准性问题,各种搜索算法、模式识别及智能方法在计算机博弈中都可以得到广泛的应用。在长时间的研究中,涌现出大量令人震惊的成果,1997年“深蓝”战胜卡斯帕罗夫的比赛就在全世界范围内引发了震动。其他很多棋类的计算机水平都已达到了世界冠军的水平。 目前,对于像五子棋、中国象棋等棋类游戏的计算机博弈算法研究已相对成熟,六子棋作为一个刚刚兴起不久的棋类游戏,其计算机博弈算法的研究还相对较少。即使目前已经出现六子棋的论坛以及比赛的平台,但只限于人人对弈。真正对于六子棋计算机博弈算法以及系统的研究还不多。六子棋的发明者台湾吴毅成教授给出了六子棋的公平性问题以及基于迫著(Threats)的胜利策略,但是对于其计算机博弈问题没有给出更加深刻的阐述,同时也没有全面解决六子棋计算机博弈问题。本文正是对六子棋计算机博弈技术的一个探索。 本文中提出的六子棋计算机博弈系统可以分为四个主要部分:搜索引擎、走法生成、评估函数和开局库。搜索引擎模块包含了比较成熟的搜索算法,以及对他们的结合和优化;走法生成模块是对搜索的结果进行比较处理,确定当前的走法;评估函数模块中本文根据棋型特征构建了着子棋力的评估函数并提出用遗传算法来做评估函数参数的调整与优化的方法;开局库存储了大量的专家棋谱,可以避免在开局时由于搜索深度的不足而带来战略上的失误,同时大大提高了对战的效率。 最后本文对六子棋计算机博弈系统进行了测试与评价,包括评估函数的准确度、搜索算法的效率以及系统的整体性能。
[Abstract]:Computer game is one of the most important and challenging research directions in the field of artificial intelligence, and its research has brought many important methods and theories for artificial intelligence. Has produced the widespread social influence and the academic influence as well as the massive research result. In the past half century, scholars around the world have spent a lot of effort on computer games, including Othello, checker, chess, Chinese chess, Gobang, go. This is because computer game is a touchstone of artificial intelligence, but board game is a standard problem of computer game. All kinds of search algorithms, pattern recognition and intelligent methods can be widely used in computer game. After a long period of research, a number of shocking results emerged, and the 1997 Deep Blue victory over Kasparov shocked the world. The computer level of many other chess pieces has reached the level of world champion. At present, the research on computer game algorithms for chess games such as Gobang and Chinese chess has been relatively mature. As a new chess game, the research on computer game algorithm of six sub-chess is relatively few. Even if there are six chess forums and game platforms, but only everyone game. There is not much research on the computer game algorithm and system of six sub-chess. Professor Wu Yicheng of Taiwan, the inventor of six sub-chess, presented the fairness of six sub-chess and the victory strategy based on forced (Threats), but did not give a more profound explanation of the computer game. At the same time, there is no overall solution to the six sub-chess computer game problem. This paper is an exploration of computer game technology of six sub-chess. The computer game system proposed in this paper can be divided into four main parts: search engine, running method generation, evaluation function and opening library. Search engine module contains more mature search algorithms, as well as their combination and optimization, walk generation module is to compare the results of the search, determine the current method; In the evaluation function module, the evaluation function of sub-chess force is constructed according to the characteristics of chess type, and a genetic algorithm is proposed to adjust and optimize the parameters of evaluation function, and a large number of expert chess spectra are stored in the opening library. It can avoid the strategic mistakes caused by the lack of search depth at the beginning, and greatly improve the efficiency of the battle. Finally, the computer game system of six sub-chess is tested and evaluated, including the accuracy of the evaluation function, the efficiency of the search algorithm and the overall performance of the system.
【学位授予单位】:重庆大学
【学位级别】:硕士
【学位授予年份】:2007
【分类号】:TP18;TP311.52

【引证文献】

相关期刊论文 前1条

1 周菁菁;;六子棋——人工智能系统的设计与研究[J];湖北广播电视大学学报;2011年10期

相关博士学位论文 前1条

1 徐长明;基于连珠模式的六子棋机器博弈关键技术研究[D];东北大学;2010年

相关硕士学位论文 前5条

1 姜勇;五子棋人机对战系统设计[D];电子科技大学;2010年

2 王林;基于有限自动机的机器博弈系统建模与搜索算法的研究[D];东北大学;2009年

3 李新星;六子棋中基于BP-TD学习的局面估值方法研究[D];东北大学;2009年

4 闵文杰;六子棋计算机博弈关键技术研究[D];重庆交通大学;2010年

5 李翠珠;六子棋计算机博弈系统的研究与实现[D];重庆理工大学;2010年



本文编号:2280664


论文下载
论文发表
教材专著
专利申请


    下载步骤:1.微信扫码,备注编号 2280664. 2.自助下载


    本文链接:http://www.bigengculture.com/kejilunwen/rengongzhinen/2280664.html