基于模糊证据推理及改进TOPSIS的多属性群决策方法  

Novel Approach for Multi-criteria Group Decision Making Based on Fuzzy Evidential Reasoning and Improved TOPSIS Method

作　　者：李少年[1] 吴良刚[1] LI Shao-nian WU Liang-gang(School of Business, Central South University, Changsha 410083, China)

机构地区：[1]中南大学商学院,湖南长沙410083

出　　处：《运筹与管理》2017年第6期16-23,共8页Operations Research and Management Science

基　　金：国家自然科学基金委创新群体资助项目(70921001);中国移动通信集团业务支撑重点联合研发项目(2015-LH-21)

摘　　要：针对不同识别框架多属性群决策问题属性准则度量的不确定性、随机性,定义基于梯形模糊数表征的属性准则评价等级相似度量,求解专家决策权重的最优解。对公共识别框架备选方案属性准则采用模糊证据推理过程综合专家评价等级置信度信息;利用可严格区分属性准则评价等级的相似度量,改进TOPSIS方法中备选方案属性准则评价等级置信度距离因子,获取备选方案逼近正负理想解的贴近度。实例分析以某通信企业电信产品市场竞争力评估为例,说明基于模糊证据推理、改进TOPSIS的多属性群决策问题求解过程,从属性准则专家模糊评价等级置信度集中获取直观的待评估产品市场竞争力排序结果,验证该方法解决此类决策问题的可行性与有效性。To deal with vague and random criteria measurement of multi-criteria group decision making （MCG- DM） among various discernment frameworks, the paper defines a similarity measurement of the assessment grades about criteria based on trapezoidal fuzzy numbers for the sake of the optimal solutions to experts＇ decision weights. Fuzzy evidential reasoning（FER）is applied to aggregate experts＇ confidential degree information of the assessment grades about criteria of the alternatives in the common discernment frameworks. The similarity meas- urement is facilitated with the capacity of discerning diverse assessment grades about criteria rigorously, which can improve the feasibility of confidential degree＇ s distance factors subjected to the assessment grades about cri- teria of the alternatives. Then, the closeness degree deduced from the technique for order preference by similarity to ideal solution（ TOPSIS） is obtained. Experiment analysis of the marketing competitiveness evaluation of tele- com products proposed by a certain communication enterprise shows the procedure of MCGDM based on FER and the improved TOPSIS method. The visual marketing competitiveness ranking results of the evaluated telecom products are acquired from experts＇ vague confidential degree information of the assessment grades about crite- ria, which verify the applicability and effectiveness of the approach proposed in the paper to tackle such decision problems.

关 键 词：多属性群决策 梯形模糊数 评价等级相似度量 置信度距离因子 市场竞争力评估 

分 类 号：C931.9[经济管理—管理学] TP391[自动化与计算机技术—计算机应用技术]