文献类型：期刊文献

英文题名：A multi-agent complex network algorithm for multi-objective optimization

作者：Li, Xueyan[1];Zhang, Hankun[2]

通讯作者：Li, XY[1]

机构：[1]Beijing Union Univ, Sch Management, Beijing 100101, Peoples R China;[2]Beijing Technol & Business Univ, Sch Business, Beijing 102488, Peoples R China

第一机构：北京联合大学管理学院

通讯机构：[1]corresponding author), Beijing Union Univ, Sch Management, Beijing 100101, Peoples R China.|[1141755]北京联合大学管理学院;[11417]北京联合大学;

年份：2020

卷号：50

期号：9

起止页码：2690-2717

外文期刊名：APPLIED INTELLIGENCE

收录：;EI(收录号：20201508398635);Scopus(收录号：2-s2.0-85082856604);WOS:【SCI-EXPANDED(收录号:WOS:000520906000001)】；

基金：This work was supported by project of science and technology research and development plan of China National Railway Group Co., Ltd. (grant numbers: K2019Z006), Beijing Social Science Foundation (grant numbers: 19ZDA05 and 19YJB016).

语种：英文

外文关键词：Multi-objective optimization; Cellular genetic algorithm; Complex network; Private archive; "Local-global" genetic operator

摘要：To deal with the multi-objective optimization problems (MOPs), this study proposes a new Multi-Objective Multi-Agent Complex Network Optimization Algorithm called MOMCNA based on the idea of Cellular genetic algorithm (CGA) and the Multi-agent complex network. Compared with the traditional CGA for multi objective problem, the individuals in the population of MOMCNA have more features of intelligent agent, the new form of neighborhood for the population, private archive for individuals, the new strategy of "local-global" genetic operator and the chaotic mutation are proposed in the new algorithm to balance the convergence and diversity of the algorithm. Seventeen unconstrained multi-objective optimization problems and seven many-objective problems are introduced and tested to evaluate the new algorithm, in addition, the classical traffic assignment problem based on different system optimum principle is also established to evaluate the new algorithm. The comparison between MOMCNA and other classical algorithms shows that the proposed MOMCNA proves to be competitive in dealing with multi-objective and many-objective optimization problems and the structure of the complex network made up of population also has effect on algorithm's performance.