基于专家系统的中国象棋残局推理研究

发布时间：2018-05-30 05:17

本文选题：专家系统 + 象棋残局　；参考：《燕山大学》2013年硕士论文

【摘要】：中国象棋一直是计算机博弈研究的热点和难点之一，在人工智能及人机交互等诸多领域有广泛的应用。在目前的中国象棋研究中，对开局和中局的研究都达到了一定程度，比较而言，对象棋残局的研究还比较薄弱和缺少，，需要不断加强，本文试图用专家系统的思想方法对中国象棋残局进行研究，即根据某些专家凭借经验所给出的残局的局面状况推理出随机出现的满足推理模式的某一残局的胜负情况。首先，说明了如何获取与表示中国象棋残局知识，主要从棋子的类型以及棋子的价值两方面进行考虑，并采用了产生式方法来表示残局规则，且给出了每条规则的可信度，然后在大量残局事实的基础上构建了中国象棋残局知识库。其次，重点阐述了推理机的实现过程，通过对残局特点进行分析，对推理机制进行描述，确定了推理的方向、匹配的算法以及冲突消解等控制策略，并给出了推理结果的可信度及其计算方法。最后，通过实验对基于专家系统的象棋残局推理方法进行验证并对整个系统的效率进行分析。首先给出了系统所用的实验平台，然后分别从规则库文件的转化和推理机的实现两个部分出发，结合前面提出的求解思想以及算法等内容，对推理过程进行实现并给出了部分实验代码，并通过具体的残局实例对系统进行了验证。最后论述了残局知识库和残局推理在计算机博弈中的应用，主要从残局数据库、搜索引擎和棋局评估这三个方面进行研究，探讨了当前情况下残局处理所面临的难点和挑战，并提出了需要改进的方向。
[Abstract]:Chinese chess has been one of the hotspots and difficulties in computer game research, and has been widely used in many fields such as artificial intelligence and human-computer interaction. In the current study of Chinese chess, the research on the beginning and the middle bureau has reached a certain degree. Comparatively speaking, the study on the chess endgame is still relatively weak and lacking, and needs to be continuously strengthened. This paper attempts to use the thought method of expert system to study the Chinese chess endgame, that is, according to the situation given by some experts with experience, we can infer the winning or losing situation of a certain final game which meets the reasoning mode at random. First of all, it explains how to acquire and represent the knowledge of Chinese chess game, mainly considering the type of chess pieces and the value of pieces, and adopts the production method to express the residual rules, and gives the credibility of each rule. Then, on the basis of a large number of residual facts, the Chinese chess residual knowledge base is constructed. Secondly, the realization process of the inference engine is expounded. Through the analysis of the characteristics of the residual, the reasoning mechanism is described, the direction of reasoning, the matching algorithm and the conflict resolution control strategy are determined. The reliability of the reasoning result and its calculation method are also given. Finally, the method of chess residual reasoning based on expert system is verified by experiments and the efficiency of the whole system is analyzed. Firstly, the experimental platform used in the system is given, and then, starting from the transformation of the rule base file and the realization of the inference machine, the paper combines the idea of solving and the algorithm, and so on. The reasoning process is implemented and some experimental codes are given, and the system is verified by a concrete residual example. Finally, this paper discusses the application of residual knowledge base and residual reasoning in computer game, mainly from the three aspects of residual database, search engine and chess evaluation, and discusses the difficulties and challenges in dealing with the residual in the current situation. The direction of improvement is also put forward.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP182

【参考文献】