摘要
多属性群体决策中,因决策问题复杂、群体成员众多难以形成一致的决策,对群体成员进行聚类形成子群体以减少群体集结复杂度,这是群体决策的重要基础。通过聚类算法提高群决策的效率与质量,首先,结合皮尔逊相关系数和切比雪夫距离定义了新的相似性度量;其次,利用模拟退火算法的参数全局寻优特性改进近邻传播算法;最后,通过实例应用与对比分析,说明新的相似性度量的优势并展示了改进算法的有效性和稳定性。
In multi-attribute group decision making, it is difficult to form consistent decisions due to the complexity of decision making and the number of group members. Preference clustering of group members to form sub-groups to reduce the complexity of group assembly is an important basis for group decision making. To improve the efficiency and quality of group decision making by clustering algorithm, firstly, a new similarity measure is defined by combining Pearson correlation coefficient and Chebyshev distance;secondly, the affinity propagation algorithm is improved by using the parameter global optimization feature of simulated annealing algorithm;finally, the advantages of the new similarity measure are illustrated and the effectiveness and stability of the improved algorithm is demonstrated by example application and comparative analysis.
作者
张煜
王磊
俞璐
马昊
郁楠
Zhang Yu;Wang Lei;Yu Lu;Ma Hao;Yu Nan(College of Communications Engineering,Army Engineering University of PLA,Nanjing,Jiangsu 210007,China;Unit 31121 of PLA)
出处
《计算机时代》
2022年第10期45-50,共6页
Computer Era
基金
国家自然科学基金资助项目(61702543,61971439)
江苏省自然科学基金资助项目(BK20191329)
中国博士后科学基金资助项目(2019T120987)
陆军工程大学基础前沿创新项目(KUYTYJQZL1906)。
关键词
群体决策
偏好聚类
相似性度量
近邻传播
模拟退火
group decision making
preference clustering
similarity measure
affinity propagation
simulated annealing
作者简介
张煜(1985-),男,汉族,福建福州人,工程师,硕士研究生,主要研究方向:系统工程、多属性决策;通讯作者:王磊(1983-),男,汉族,江苏丹阳人,副教授,硕士生导师,博士,主要研究方向:系统仿真、优化理论。
