针对当前的信息集结算子研究在解决协同设计方案多准则群体决策中的不足, 提出一种基于密度峰值法的类信息可靠性诱导有序加权平均算子. 首先由决策者根据模糊数学的隶属度理论对备选设计方案进行定量评价, 形成决策矩阵,在此基础上将决策者对每个设计方案的决策结果表示为特征向量; 然后提出一种改进的密度峰值聚类算法对每个设计方案的决策向量进行聚类, 并获取每一类的聚类中心作为代表数据; 最后计算每类决策信息的可靠性, 利用类信息可靠性诱导的有序加权平均算子进行信息集结, 得到每个设计方案的综合评价结果. 以某型号卫星的结构设计方案决策为例, 验证了文中方法在解决多准则群体决策问题中的有效性.%Aiming at the shortcomings of the current information aggregation operators in solving the mul-ti-criteria group decision making problems of collaborative design alternatives, this paper proposed a class-reliability-induced OWA operator based on density peak clustering algorithm. Firstly, the decision makers adopted the membership theory of fuzzy mathematics to quantitatively evaluate the alternative design scheme and form the decision matrix. The decision-making results of each decision maker for each scheme were taken as feature vectors. Secondly, an improved density peak clustering algorithm was proposed to cluster the deci-sion vectors of each design scheme and obtain the clustering center as representative data of each cluster. Thirdly, the reliability of each cluster was calculated, and the information was aggregated using the class-reliability-induced OWA operators. The comprehensive evaluation results of each design alternative were obtained. The collaborative decision of a certain type of satellite structure design was taken as an example to demonstrate the effectiveness of proposed approach in solving the multi-criteria group decision making prob-lems.
展开▼
