三种新的基于相似性的加权模糊推理方法  

Three New Kinds of Similarity-Based Weighted Fuzzy Reasoning Methods

作　　者：哈明虎[1] 李海军[1] 

机构地区：[1]河北大学数学与计算机学院,河北保定071002

出　　处：《计算机工程与应用》2006年第28期34-37,共4页Computer Engineering and Applications

基　　金：国家自然科学基金资助项目(编号:60473045;60573069);河北省软科学资助项目(编号:044572150);河北省教育厅科学研究计划重点资助项目(编号:2005001D)

摘　　要：在模糊专家系统中,模糊推理方法的优劣是衡量系统性能好坏的关键指标。基于相似性的加权模糊推理是针对模糊信息发展的一种既简单又灵活的方法,其关键是模糊集合相似度的定义。在文中,提出了三种新的相似度的定义,讨论了与其相应的模糊推理方法;通过实例验证了三种新方法的有效性,并与已有的几种方法进行了比较分析。In fuzzy expert systems,the performance of fuzzy reasoning methods is an important factor related to the capability of the system.Similarity-based weighted fuzzy reasoning is a kind of simple and flexible method that was developed when fuzzy information was introduced,the key aspect of which is the definition of similarity measure of fuzzy sets.In this paper,three kind of similarity measures are proposed and the corresponding fuzzy reasoning methods are discussed.We calculate several examples to show the validity of these methods and compare them with other existing similarity-based reasoning methods.

关 键 词：加权模糊产生式规则 相似度 模糊推理方法 专家系统 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]