信息不完全确定的群体多准则语言聚类方法  

Group multi-criteria linguistic clustering method with incomplete certain information on weights

作　　者：王坚强[1] 王君[1] 孙超[1] 

机构地区：[1]中南大学商学院,湖南长沙410083

出　　处：《系统工程与电子技术》2007年第11期1872-1876,共5页Systems Engineering and Electronics

基　　金：湖南省社会科学基金(05YB74);湖南省软科学计划项目(06FJ4126);湖南省哲学社会科学评审委员会项目(0608064A)资助课题

摘　　要：针对决策者权重和准则权重信息不完全确定,且评价信息为语言值的多准则聚类问题,提出了一种新的求解方法。首先通过证据推理算法,将方案集的群体多准则语言评价信息转化为方案集模式样本,该模式样本是决策者权重和准则权重的不确定信息的非线性函数,然后根据K-均值方法,利用惯性粒子群算子对模式样本集进行聚类,得到最优聚类中心和方案集的最优聚类。最后实例说明该方法的可行性和有效性。For group multi-criteria clustering problems with linguistic assessment information, a new method is proposed, in which the decision-makers＇ weights are incomplete certain, and so are criteria weights. Firstly, the group linguistic assessment information of alternatives is transformed into a model sample of alternatives, Then the model sample of alternatives is aggregated with inertial particle swarm optimization operator. The nonlinear mixed integral programming model is constructed based on the incomplete certain information of the decision makers＇ weights and criteria weights, the programming model is worked out with the improved inertial weights＇ particle swarm optimization algorithm, and the optimal clustering central and optimal clusters are obtained. Finally, an example is given to show the feasibility and availability of this method.

关 键 词：群体多准则聚类 语言评价 信息不完全确定 二元语义 粒子群算法 

分 类 号：N945.25[自然科学总论—系统科学]