基于灰关联深度系数的评价指标客观权重极大熵配置模型  

Maximum entropy configuration model of objective index weight based on grey correlation deep coefficient

作　　者：汪群峰[1] 金佳佳[1] 米传民[1] 方志耕[1] 

机构地区：[1]南京航空航天大学经济与管理学院,南京211106

出　　处：《控制与决策》2013年第2期235-240,共6页Control and Decision

基　　金：国家社科重大项目(10zd-014);国家社科重点项目(08AJY024);国家自然科学基金面上项目(70971064);国家自然科学基金青年项目(70901041;71002046);国家自然科学基金项目(71171112);中国博士后基金项目(20090450588);中央高校基本科研业务费专项资金项目(NS2012029)

摘　　要：在多属性决策领域中,经典评价指标客观赋权的极大熵模型主要依据指标评价值的差异度进行赋权,极易导致对评价值差异度小的指标作出不重要(权值小)的错误判断,从而产生决策偏差.对此,首先提出灰关联深度系数的概念,以表征指标客观权重包含的信息量大小;同时运用极大熵理论建立评价指标客观权重的极大熵配置模型,以确定多属性决策指标权重,较好地解决了经典模型存在的严重缺陷;最后通过实际案例的对比分析,验证了所提出解决方案的优良性能.In the field of multiple attribute decision making,the classic weighing values for objective index using maximum entropy model are mainly based on the difference of the value of evaluation indexes,which easily results in the wrong judgment that some index is not important（small index weight） because of the small difference degree,so that the decision deviation is produced.The grey correlation deep coefficient is proposed to characterize the information size contained by objective index weight.And the maximum entropy optimization model of objective weight is established by combining with the maximum entropy criterion to compute the objective weight of multiple attribute decision,which solves the classic model’s serious defect better.Finally,the comparison analysis of the actual cases demonstrates the good performance of the proposed method.

关 键 词：灰色关联分析 灰关联深度系数 极大熵 多属性决策 指标权重 

分 类 号：C934[经济管理—管理学]