摘要
针对群体成员偏好信息以效用值形式给出的大群体决策问题,提出了判断群体成员提供信息量多寡程度的熵权方法,去除提供较少信息量的成员,形成群体关于决策方案的效用矩阵.利用聚类方法对大群体成员效用向量进行聚类,根据聚类结果确定成员权重,将该权重与效用矩阵合成获得决策方案排序向量.提出了成员意见反映度指标和差异度指标,对群决策结果进行评价.最后通过一个实例说明该方法的有效性和实用性.
Aiming at the problem of large group decision-making based on utility valued preference information, the entropy weight method is presented to appraise decision members＇ information, delete these members who provide less information and form the group utility matrix about decision schemes. These members are classified by using the clustering method. Members＇ weights are confirmed by using the clustering result. And the ordering vectors of decision schemes are obtained by synthesizing the members＇ weight vectors and utility matrixes. The index of members＇ opinion reflection degree and the difference degree is proposed to evaluate the result of group decisionmaking. Finally, a real example showsthe validity and practicability of the method.
出处
《控制与决策》
EI
CSCD
北大核心
2009年第3期440-445,450,共7页
Control and Decision
基金
国家自然科学基金重点项目(70631004)
国家自然科学基金项目(70871121)
关键词
大群体
群决策
效用值
反映度
差异度
Large group
Group decision
Utility value
Reflection degree
Difference degree
