动态环境下的群组专家多准则变权决策方法.pdfVIP

ComputerEngineeringandApplications计算机工程与应用 2015，5l(2) l @博士论 坛@ 动态环境下的群组专家多准则变权决策方法 孙永河 ，段万春 ，李春好 ，许成磊 SUN Yonghe，DUAN W anchun，LIChunhao，XU Chenglei 1．昆明理工大学 管理与经济学院，昆明 650093 2．吉林大学 管理学院，长春 130025 1．FacultyofManagementandEconomics，KunmingUniversityofScienceandTechnology,Kunming650093，China 2．SchoolofManagement，JilinUniversiyt,Changchun130025，China SUN Yonghe，DUAN W anchun，LIChunhao，eta1．Multi-criteriagroup decision makingapproachwithvariable weightsindynamicenvironment．ComputerEngineeringandApplications，2015，51(2)：1-6． Abstract；ClassicMulti—CriteriaDecisionMaking(MCDM)methodisunabletodealwiththedynamiciyt inrealworld． Also，itiSdifficulttoreflectthefeedbackactionthatAltemativeCluster(AC)dominatesCriteriaCluster(CC)．Toover— come the above mentioned drawbacks，in thispaperamulti-criteria group decision making approach with variable weightsinadynam icenvironmentissuggestedbysynthesizingbothdataenvelopmentanalysisandanalyticnetworkpro— cess．ComparedwithclassicM CDM method，theoutstandingadvantagesfortheapproachlieinfollowingthreepoints．A dynamicevolvementmechanism forM CDM modelisproposed．A novelgroupinformationdistillingwayisgivenbyana— lyzingthesestatementsthateachalternativeliesincriteriatoexpressexpe~preference，whichensuresnolossofeveryinfor- marion．Itmayrealizethedecisionwithvariableweightsforalternatives，therefore，nonlinearinteractionrelationbetween AC andCC couldbewcllreflected．TheapproachiSvalidatedtobefeasibleandscientificandcanbewellappliedto solvetherealworlddynamicdecisionissues，suchasselectingdynamicdisasterreliefalternatives，supplierdynamiceval- uation，etc． Keywords：dynamicenvironment；groupexperts；Multi—CriteriaDecisionMaking(MCDM)；variableweights 摘 要：为克服经典多准则决策(MCDM)方法不适应动态的决策环境、难以反映方案集对准则集的非线性反馈效应 等方面缺陷，通过运用网络分析和数据包络分析技术，提 出一种动态环境下的群组专家多准则变权决策方法。较之 于经典MCDM方法，新方法主要创新之处在于：给出了MCDM模型的动态演化机理；通过专家对方案所处准则状 态予以有偏好 (无偏好)判断，提 出一种保证信息无损的群组专家信息提取方式；实现了对方案的变权评价，有效反 映出蕴含在系统 内部的准则集与方案集的非线性交互作用关系。实例验证结