运筹与管理 ›› 2015, Vol. 24 ›› Issue (5): 197-205.DOI: 10.12005/orms.2015.0177

• 应用研究 • 上一篇    下一篇

基于灰类敏感度系数的评价指标客观权重
极大熵配置模型

刘红旗1~3,方志耕1,2李维东1,2陶良彦1,2   

  1. 1.南京航空航天大学经济与管理学院,江苏南京210016;
    2.南京航空航天大学科学发展研究中心;
    3.南京航空航天大学纪委办
  • 收稿日期:2013-12-17 出版日期:2015-10-12
  • 作者简介:刘红旗(1981-),男,博士研究生,纪委办主任,助理研究员,研究方向:灰色系统理论、评价理论与方法;方志耕(1962-),男,教授,博导,主要从事灰色系统、复杂装备研制与管理、质量管理等研究;李维东(1988- ),男,硕士,专业: 统计学,方向:质量管理;陶良彦(1988- ), 男,硕士,专业: 管理科学与工程,方向:复杂装备研制。
  • 基金资助:
    国家自然科学基金资助项目(70971064,71173106,71171113);国家社科基金重大项目(10zd&014);国家社科基金重点项目(12AZD102);教育部人文社科青年基金项目(12YJC630115,12YJC630276);中央高校基本科研业务费专项科研项目资助(NJ20130021,NJ20140032);江苏高校哲学社会科学重点研究基地重大项目(2010JDXM014,2010JDXM015,2012JDXM003)

The Maximum Entropy Configuration Model of Objective Index Weight
Based on Grey Class Sensitivity Coefficient

LIU Hong-qi1~3, FANG Zhi-geng1,2, LI Wei-dong1,2, TAO Liang-yan1,2   

  1. 1.College of Economics and Management, Nanjing University of Aeronautics and Astronautics,Nanjing210016,China;
    2.Scientific Development Research Center, Nanjing University of Aeronautics and Astronautics;
    3.Discipline Inspection Commission Office, Nanjing University of Aeronautics and Astronautics
  • Received:2013-12-17 Online:2015-10-12

摘要: 评价指标权重的确定是多属性决策问题中至关重要的环节。然而,既有研究利用评价指标值之间的差异性进行指标的客观赋权,往往忽略了评价指标对被评价对象全体及所在系统的重要性。本文基于事物自然本质属性差异,运用灰色关联聚类将评价对象划分到预设的类别;借鉴有无对比分析法的思想,定义了反映指标对全体被评价对象类别影响程度的灰类敏感度系数;根据极大熵准则,建立了基于灰类敏感度系数的客观权重极大熵配置模型,以确定多属性决策指标的权重。并通过与文献[21,26]中实际案例的对比分析,说明了本模型的有效性与更贴近现实性,为解决多属性决策指标客观赋权问题提出了一个新思路。

关键词: 多属性决策, 灰色关联聚类, 灰类敏感度系数, 极大熵, 指标权重

Abstract: To determine the weight values of assessment indexes is the crucial link in multiple attribute decision making problems. However, today’s researches have used the evaluation index difference, which tends to ignore the evaluation index of the importance for the evaluation object and the system. Based on the natural essence attribute differences, this article uses the grey correlation clustering to classify the evaluation object into the default category, and by the analysis method of being with or without contrast, defines the grey class sensitivity coefficient which influences the impact of the indexes for the evaluation of all the object classes; and based on the maximum entropy criterion,sets up the objective weight maximum entropy configuration model based on the grey class sensitivity coefficient to determine the weight of multiple attribute decision making. Through the contrast analysis of actual cases in ref. [21,26], it shows the effectiveness and closeness to the reality of this model, which has put forward a new thought for solving multiple attribute decision making problem.

Key words: multiple attribute decision making, grey correlation clustering, grey class sensitivity coefficient, maximum entropy, objective index weight

中图分类号: