摘要
针对决策者权重与属性权重完全未知的区间直觉模糊多属性群决策问题,给出一种基于相关系数及改进TOPSIS法的多属性群决策方法。将各决策者同等对待,得到各方案关于每个属性的评价均值,由各决策者在每个方案下关于单个属性的区间直觉模糊评价值与其评价均值的相关系数,获取在单个属性下体现出的各决策者权重。基于各决策者权重得到群体区间直觉模糊决策矩阵,构建各方案与正理想方案加权相关系数总和最大化(或与负理性方案加权相关系数总和最小化)的目标规划模型确定各属性权重。以两组属性权重向量分别得到各方案与正、负理想方案的加权相关系数,依据改进的TOPSIS法计算各方案与正理想方案的相对相关系数,并以此得到各方案的优先序。投资项目选择算例说明该群决策方法有效性与合理性。
With respect to multi-attribute group decision making problem with interval-valued intuitionistic fuzzy numbers, where the weights of decision-makers and attributes are completely unknown. All decision-makers are treated equal, the mean evaluation of alternatives under every attributes are obtained. Then, the correlation coefficients between individual attribute values and its mean evaluation are calculated, decision-makers weights are determined by single attribute under every alternative. Group interval-valued intuitionistic fuzzy decision matrix is proposed based on decision-makers weights, a target programming model is established based on the sum total maximum of correlation coefficients between each alternative and the positive ideal alternative （or based on the sum total minimum of the correlation coefficients between each alternative and the negative ideal alternative）, the weights of attributes can be get by the model. The weighted correlation coefficient between each alternative and positive （negative） ideal alternative is proposed by two attribute＇s weight vectors, relative correlation coefficient between each alternative and the positive ideal alternative is calculated based on improved TOPSIS method, the ranking order of all alternatives is proposed. Finally, investment project selection example is given to show the effectiveness and feasibility of the proposed method.
出处
《模糊系统与数学》
CSCD
北大核心
2016年第5期132-141,共10页
Fuzzy Systems and Mathematics
基金
国家自然科学基金项目(71272049
71402142)
高等学校博士学科点专项科研基金项目(20126102110052)
西北工业大学人文社科与管理研究基金项目(3102014RW0008)
关键词
多属性群决策
区间直觉模糊数
相关系数
TOPSIS法
multiple attribute group decision making
interval-valued intuitionistic fuzzy number
correlation coefficient
TOPSIS method
