基于证据距离的改进DS/AHP多属性群决策方法  

An improved method of aggregation in DS/AHP for multi-criteria group decision-making based on distance measure

作　　者：姚爽[1] 郭亚军[1] 黄玮强[1] 

机构地区：[1]东北大学工商管理学院,沈阳110819

出　　处：《控制与决策》2010年第6期894-898,共5页Control and Decision

基　　金：辽宁省财政科研基金项目(2007B002)

摘　　要：证据推理/层次分析(DS/AHP)方法采用了AHP法的层次结构模型和证据理论的分析过程,为不确定多属性决策问题的解决提供了新思路,但其在构造知识矩阵中用"0"代表残缺信息是不合理的.鉴于此,对DS/AHP方法进行了改进,并将改进后的方法拓展到群决策中,研究了专家群体权向量的确定方法.具体地,引入证据距离的概念,通过计算专家证据的综合距离来对专家赋权,体现了群决策中的多数人规则.As a nascent method of multi-criteria decision-making with uncertainty,DS/AHP is structured similar to the analytic hierarchy process（AHP） and its analytic processes are based on Dempster-Shafer theory.While it is unsuitable to make＂0＂stand for the missing information during the construction of knowledge matrix in DS/AHP.Therefore,some improvement work is done for DS/AHP.The improved method is used in group decision-making to study the method of fixing the weighting vector for individual members of a decision-making group.Particularly,by regarding the collective distance measure as induced variable,the weighting vector of group members is given and the＂majority rule＂of group decision-making is used.

关 键 词：群决策 证据推理/层次分析 证据距离 

分 类 号：C934[经济管理—管理学]