改进灰色证据理论的区间数决策方法研究  

Research on interval numbers decision-making based on improved grey incidence and D-S evidence of theory

作　　者：李存斌[1] 赵坤[1,2] 苑嘉航 

机构地区：[1]华北电力大学经济与管理学院,北京102206 [2]国华能源投资有限公司,北京100007

出　　处：《天津工业大学学报》2014年第4期70-75,共6页Journal of Tiangong University

基　　金：国家自然科学基金项目(71271084);国家电网公司科技项目(521820140017)

摘　　要：针对指标值以区间数形式给出的多指标决策问题,提出一种改进的灰色关联与D-S证据理论结合的决策方法.为避免造成指标不确信度的偏离,本文应用灰色区间数距离来计算指标不确信度.在此基础上,构建多指标下备选方案的Mass函数,通过Dempster合成法则进行信息融合,根据置信函数大小对方案进行排序.算例分析结果表明:本文方法进一步降低了决策的初始不确定性和融合后的整体不确定性,扩大了不同方案置信函数之间的差距,使各方案的优劣性更加明显,有利于决策者做出选择.As for the multiple attribute decision making, in which the attribute values are in the form of interval numbers, a method based on improved grey incidence and D-S theory evidence is proposed. In order to avoid deviating accuracy, the distance of grey interval number is used to analyze the uncertain degrees of different indices . Based on this, the Mass functions of different projects in different indices are obtained, and the rule of Dempster combination is used to fuse information, and the belief function values are ordered to sort the projects. The example analysis show that, the uncertainty of decision making can be decreased further, and also the gap among the belief function value of different projects has been enlarged, which makes the projects′ advantages and disadvantages more obvious and will be more easier for decision making.

关 键 词：灰色关联法 D-S证据理论 区间数 决策 

分 类 号：N941.5[自然科学总论—系统科学]