基于极大熵理论的评价指标组合赋权模型  

Combination Weight Model of Evaluation Index Based on the Maximum Entropy Theory

作　　者：丁元明[1] 邱磊 贾良陈[1] 

机构地区：[1]辽宁省通信网络与信息处理重点实验室大连大学信息工程学院,辽宁大连116622 [2]中国船舶重工集团第701研究所规划处,湖北武汉430064

出　　处：《数学的实践与认识》2014年第4期16-22,共7页Mathematics in Practice and Theory

基　　金：辽宁省教育厅科学研究项目(LS2012440);创新团队项目(2009T004)

摘　　要：针对目前指标组合赋权方法中存在的赋权过于"僵硬"的通病,基于Jaynes E T提出的最大信息熵原理,构建了一种新的指标赋权模型.在模型中,以客观因素为目标规划主体,将主观因素转化为目标规划的约束条件,实现了主客观条件的融合.实例分析表明方法准确、有效,具有较高的应用价值.Aiming at a existing over-generalization of too much ＂stiff＇ in the combination weight method of index,A new kind of index weighting model is constructed based on the maximum entropy which proposed by E T Jaynes.In this model,as objective factors for the goal programming subject,translate subjective factors into the constraints of the goal programming,which received a good combination of subjective and objective conditions.The analysis of the instance demonstrates that the method with accuracy and efficiency,so its application has higher values.

关 键 词：组合权重 极大熵 目标规划 效能评估 

分 类 号：O221[理学—运筹学与控制论]