基于决策变量分组的粒子群算法求解大规模优化问题  

Particle swarm optimization algorithm based on grouping of decision variables to solve large-scale optimization problems

作　　者：白晓慧 何小娟[1] 孙超利[2] 时振涛 张国晨 BAI Xiaohui;HE Xiaojuan;SUN Chaoli;SHI Zhentao;ZHANG Guochen(School of Applied Science,Taiyuan University of Science and Technology,Taiyuan Shanxi 030024;School of Computer Science and Technology,Taiyuan University of Science and Technology, Taiyuan Shanxi 030024)

机构地区：[1]太原科技大学应用科学学院,山西太原030024 [2]太原科技大学计算机科学与技术学院,山西太原030024

出　　处：《宁夏师范学院学报》2020年第4期50-56,共7页Journal of Ningxia Normal University

基　　金：国家自然科学基金(61876123);山西省自然科学基金(201801D121131);山西省自然科学基金(201901D111264);山西省优秀人才科技创新项目(201805D211028);山西省留学回国人员科技活动择优资助项目.

摘　　要：针对社会学习粒子群算法在求解大规模优化问题时存在的收敛速度慢以及种群多样性缺失等问题,提出一种基于决策变量分组的粒子群算法.根据决策变量间的相关性对决策变量分组,提高算法的收敛速度.采用反向学习策略,通过生成反向解,提高算法的全局寻优能力.采用CEC2010测试函数集对本文算法进行测试,仿真结果与已有典型算法进行对比,验证了本文算法的有效性.Aiming at the problems of slow convergence speed and lack of population diversity of social learning particle swarm optimization algorithm when solving large -scale optimization problems,a particle swarm algorithm based on grouping decision variables is proposed.The decision variables are grouped according to the correlation between the decision variables,and the convergence speed of the algorithm is improved.The opposition learning strategy is used to improve the global optimization ability of the algorithm by generating opposition solutions.The CEC2010 test function set is used to test the proposed algorithm,and the simulation results are compared with existing typical algorithms to verify the effectiveness of the proposed algorithm.

关 键 词：决策变量分组 反向学习 大规模优化问题 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]