模式匹配算法是规则软件系统的核心技术.本文叙述规则软件系统的基本概念、工作原理和主要推理方式;研究规则软件系统的模式匹配问题,总结了改善模式匹配效率的方法和技术;介绍Rete、Treat、Leaps和Matchbox四种典型的前向推理顺序模式匹配算法,分析研究了各个算法使用的提高匹配效率的技术及其优缺点,并对模式匹配算法当前研究进展和规则软件系统的应用现状进行了阐述.
Pattern-matching algorithm is the core technique of rule-based systems.This paper briefly introduces the basic conception,working mechanism and the main inference methods of the rule-based systems.The pattern-matching problem of rule-based systems is studied and the techniques and methods to improve the performance of the pattern-matching algorithm are described.Four typical forward chaining sequence pattern-matching algorithms including Rete,Treat,Leaps and Matchbox are stated.The methods and techniques used in each algorithm to improve the efficiency of the match process are analyzed;their advantages and disadvantages are depicted as well.Finally,the current research progress on the pattern-matching algorithm and the situation to apply rule-based system in various industries are reported.