聚散优化算法:一种新的启发式算法  

Gathering and dispersion optimization algorithm:a new heuristic algorithm

作　　者：李瑞 赵新超[1,2] 郭赛 袁健美 LI Rui;ZHAO Xinchao;GUO Sai;YUAN Jianmei(School of Science,Beijing University of Posts and Telecommunications,Beijing 100876,China;Hunan Key Laboratory for Computation and Simulation in Science and Engineering,Xiangtan University,Xiangtan 411105,China;School of Mathematics and Computational Science,Xiangtan University,Xiangtan 411105,China)

机构地区：[1]北京邮电大学理学院,北京100876 [2]湘潭大学科学工程计算与数值仿真湖南省重点实验室,湘潭湖南411105 [3]湘潭大学数学与计算科学学院,湘潭湖南411105

出　　处：《计算机集成制造系统》2020年第3期718-731,共14页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(61973042);北京市自然科学基金资助项目(1202020)。

摘　　要：为了将最优化方法与思想融入群智能优化的研究,提出了允许重复的存档和种群重置策略、聚集算子和分散算子,并分析了这些操作对算法性能的影响。在3种操作基础上,提出一种新的启发式算法——聚散优化算法(GAD)。将所提算法与4个经典启发式算法(SPSO2011,CoDE,SaDE和IGHS)相比较,用于求解20个经典的函数最优化问题。结果显示,聚散优化算法能得到较其他算法更优秀的解,从而说明了所提聚散优化算法及其操作算子的有效性。In order to integrate the optimization methods and ideas into the research of swarm intelligence,the strategies of allowing duplicate archiving and population reset,gathering operator and dispersion operator were proposed.The mathematical principles and the performance of these operators were analyzed.Based on these operations,a new heuristic algorithm,Gathering and Dispersion(GAD)optimization algorithm was proposed.Then,the proposed GAD algorithm was compared with four classical heuristic algorithmed(SPSO2011,CoDE,SaDE and IGHS)based on 20 classic benchmark.Results showed that GAD algorithm outperformed other competitors.It indicated that the proposed operations and the idea of popularizing GAD algorithm was effective.

关 键 词：聚散算法 重复存档 种群重置 聚集算子 分散算子 

分 类 号：TP278[自动化与计算机技术—检测技术与自动化装置] TP391[自动化与计算机技术—控制科学与工程]