基于广义三角模糊数的基本概率赋值构造方法  

Basic probability assignment construction method based on generalized triangular fuzzy number

作　　者：肖建于[1,2] 童敏明[1] 朱昌杰[2] 王小蕾[2] 

机构地区：[1]中国矿业大学计算机科学与技术学院,徐州221116 [2]淮北师范大学计算机科学与技术学院,淮北235000

出　　处：《仪器仪表学报》2012年第2期429-434,共6页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(61102117);国家重点基础研究发展计划(973计划)(2007CB209407)资助项目

摘　　要：Dempster组合规则的有效运用取决于相应基本概率赋值的合理建立,但是如何生成各个证据的基本概率赋值仍然是一个有待解决的问题。基于广义三角模糊数表示模型,提出了一种基本概率赋值生成方法。该方法利用样本数据的最小值、平均值和最大值构造单元素命题的三角模糊数表示模型,多子集命题的广义三角模糊数用单元素命题的三角形的交叠部分来表示。通过广义三角模糊数的隶属度来构造基本概率赋值。数值算例结果表明,提出的基本概率赋值获取方法是实用有效的。The effective application of Dempster＇ s combination rule depends on the rational construction of corre- sponding basic probability assignment （BPA）. However, how to determine basic probability assignment is still an open issue. Based on the generalized triangular fuzzy number model, a method for obtaining basic probability assignment is proposed. In this proposed method, the triangular fuzzy number model of singleton proposition is constructed using the minimum, average and maximum values of the sample data. The generalized triangular fuzzy number model of multiele- ment proposition is represented by the crossing area of the fuzzy numbers of the singleton propositions. Based on the degree of membership of the generalized triangular fuzzy number, the basic probability assignments are obtained. Nu- merical examples indicate that the proposed method for determining basic probability assignment is practicable.

关 键 词：证据理论 基本概率赋值 三角模糊数 目标识别 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]