不确定偏好序信息下多时期匹配决策的证据融合法  

AN EVIDENTIAL FUSION METHOD FOR MULTIPERIOD MATCHING DECISION-MAKING WITH UNCERTAIN PREFERENCE ORDINAL INFORMATION

作　　者：陈圣群[1,2] 王应明[1] 施海柳[1,2] 

机构地区：[1]福州大学决策科学研究所,福州350108 [2]福建江夏学院电子信息科学学院,福州350108

出　　处：《系统科学与数学》2016年第4期502-512,共11页Journal of Systems Science and Mathematical Sciences

基　　金：国家自然科学基金项目(71371053;71403055);福建省自然科学基金项目(2015J01279);福建省中青年教师教育科研项目(JA14322);福建江夏学院青年科研项目(JXZ2014003);福建江夏学院教育教学改革项目(J2015A004)资助课题

摘　　要：针对具有不确定偏好序信息的多时期匹配决策问题,提出了一种基于证据推理的决策方法.首先,给出了位置得分置信度的相关描述;然后,把匹配主体中纵向评价和横向评价的序数得分情况作为证据,并通过证据组合求出双边匹配融合度;在此基础上,构建了优化模型,获得双边匹配方案.最后,通过算例说明了该方法的可行性和有效性.A decision-making method is proposed to solve the dynamic matching decision-making problem with uncertain preference ordinals. First, the related descriptions of position score belief are given. Then, the ordinal scores of the horizontal evaluation and vertical evaluation in both matching agents are taken as evidences, and the fusion degrees of two-sided matching agents are got by combining evidences. Furthermore, an optimization model is developed to obtain two-sided matching scheme. Finally, feasibility and validity are proved through illustrative examples.

关 键 词：匹配决策 证据组合 融合度 

分 类 号：O225[理学—运筹学与控制论]