不完全信息下一种信度判断矩阵及其集结方法  

A Belief Comparison Matrix and Its Integration Method with Incomplete Information

作　　者：张洪涛[1,2] 朱卫东[1] 

机构地区：[1]合肥工业大学管理学院.安徽合肥230009 [2]安徽理工大学理学院,安徽淮南232001

出　　处：《系统工程》2010年第7期107-110,共4页Systems Engineering

基　　金：国家自然科学基金委主任基金资助项目(J0910017);安徽省高校省级自然科学研究项目(KJ2008B148)

摘　　要：针对现有判断矩阵在不完全信息环境下不能准确反映决策者的偏好,以及群决策背景下难以将专家偏好信息集结等问题,将证据推理方法引入到判断矩阵中,研究一种新型不确定判断矩阵——信度判断矩阵。在不完全信息情况下,用比较值及其信度表示两两比较时决策者的偏好,用证据推理算子将专家的意见(信度判断矩阵)进行集结,利用加权几何平均算子将合成后的信度判断矩阵转化为互反判断矩阵并求其权向量,最后结合例子来检验该方法的有效性。Decision-makers＇ preferences cannot be expressed correctly via the original comparison matrices under the incomplete information environments.In addition,information is hard to deal with when integrating multiple experts＇ preferences.To solve these problems,the evidential reasoning（ER） approach is introduced into the matrix,and a belief comparison matrix is studied.The pairwise comparison values with respect to decision-makers＇ preference are given with belief values under incomplete information circumstance.Moreover,the belief judgment matrices are integrated by the evidence reasoning combination rules.Then the integrated matrix is transformed into the reciprocal judgment matrix by a weighted geometric averaging operator,and the weights vector can be worked out by the original approach of AHP.Finally,an example is given to testify the utility of this approach.

关 键 词：层次分析法 信度判断矩阵 证据推理 不完全信息 

分 类 号：C934[经济管理—管理学]