作　　者： ;

机构地区： 华南农业大学

出　　处： 《运筹与管理》 2015年第1期122-128,共7页

摘　　要： 针对同时包含可线性补偿和不可线性补偿两种属性且属性值为确实数、区间数、语言信息的风险型多属性决策问题，提出一种基于消错理论的决策方法。首先，在消错理论的基础上将属性分为关键型、重要型和冗余型三类，结合属性值的类型分别给出对应的错误函数和极限损失值；接着，对关键型属性赋予极小权重，在保留关键型属性“一票否决”功能的同时又突出重要属性的作用；最后，根据对待错误损失的不同态度，建立计算错误损失值的三种方法，通过计算期望错误损失值对备选方案进行排序。通过新市民信息服务项目的例子，说明该方法的有效性和可行性。 In multiple attributes decision making problem under risk , the attribute has compensation features of linear and nonlinear and the attribute values have formats of crisp numbers , interval numbers and linguistic infor-mation.Thus, a new decision making approach on the base of error-eliminating theory is proposed .Firstly, by the theory of error-eliminating , decision making attributes are divided into key attribute , important attribute and redundant attribute .Error function and maximum loss values are given on the basis of these three kinds of attrib-utes and the type of attribute values .And then , minimum weight is given to key attribute , in this way the “one vote negation” function of key attribute is maintained and the important role of important attribute is highlight .At last, three methods of calculating the error-loss value are given, according to the different attitude about the error-loss, and the better strategy are prioritized and selected according to the expected error-loss values. Besides, the feasibility and effectiveness of this model have been proved by an example of new citizen information service project .

关 键 词： 决策科学 消错决策方法 消错理论 混合 风险 错误损失

分 类 号： [C934 N945.16]

领　　域： [] []