基于Vague集加权相似度量的双向近似推理  

Bidirectional Approximate Reasoning Based on Weighted Similarity Measures of Vague Sets

作　　者：王天江[1] 卢正鼎[1] 李凡[1] 

机构地区：[1]华中科技大学计算机学院,湖北武汉430074

出　　处：《计算机工程与科学》2002年第6期96-100,共5页Computer Engineering & Science

基　　金：国家高性能计算基金资助项目 (0 0 3 0 3 ) ;华中科技大学研究基金资助项目 (M990 15 )

摘　　要：本文提出了一种新的Vague集的加权相似度量方法 ,以解决文献 [1 ]中关于Vague集相似度量的某些缺陷以及权向量中各个分量难以确定的问题 ,并且提出了Vague集间相似方向的概念 ,用它来描述两个相似Vague集中哪个所包含的信息更精确 ,并给出了一个判定方法。在此基础上给出了一种基于Vague集加权相似度量的双向近似推理方法 ,该方法更好地利用了Vague集信息的精确性 ,从而提高了推理的精确性和适用性。In this paper, we introduce a new weighted similarity measure for vague sets in order to solve some faults in and get each index of a weight vector.We propose the concept of similarity direction between two vague sets to describe which one could give more accurate information. At the same time, we propose a method for determining the similarity direction. Based on the above, we present a bidirectional approximate reasoning method based on weighted similarity measures of vague sets, which fully uses the message accuracy of vague sets. Then this method improves the accuracy and applicability of approximate reasoning, and also provides a useful tool for approximate reasoning in intelligent systems.

关 键 词：FUZZY集 VAGUE集 加权相似度量 相似方向 双向近似推理 人工智能 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]