求解大规模多目标问题的改进粒子群算法  

A Modified Particle Swarm Optimization for Large-scale Multi-objective Problems

作　　者：兰丽尔 孙超利[2] 何小娟[1] 谭瑛[2] LAN Li-er;SUN Chao-li;HE Xiao-juan;TAN Ying(College of Applied Science,Taiyuan University of Science and Technology,Taiyuan 030024,China;Department of Computer Science and Technology,Taiyuan University of Science and Technology,Taiyuan 030024,China)

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

出　　处：《太原科技大学学报》2020年第4期249-256,共8页Journal of Taiyuan University of Science and Technology

基　　金：国家自然科学基金(61403272,61472269);山西省留学回国人员科技活动择优资助项目;东北大学流程工业综合自动化国家重点实验室开放课题。

摘　　要：针对大规模多目标优化问题,提出了一种基于分解的改进粒子群算法。该方法将分解策略与社会学习粒子群优化算法相结合引入到个体的学习过程中,针对每个个体及其邻域个体,计算其沿权重向量方向与参考点之间的距离以及与权重向量之间的距离并对它们进行排序,个体通过学习离参考点近的任意个体以及离权重向量近的所有个体实现位置的更新。在5个ZDT测试函数上进行了500维和1000维的测试对比,结果表明本文所提的算法具有较好的收敛性以及分布均匀性。An improved particle swarm optimization algorithm based on decomposition is proposed for large-scale multi-objective optimization problems. In the method,the social learning particle swarm optimization is proposed to incorporate into individual learning of particle swarm optimization for multi-objective optimization based on the decomposition framework. Each individual and its neighbors will calculate the distances to the reference point along the weighted vector as well as the distances to the weighted vector,and then both kinds of distances will be sorted.Individual will update its position by learning from individuals closer to the reference point as well as those closer to the corresponding weighted vectors. The experimental results on five 500-dimensional and 1000-dimensional multiobjective problems showed such method has better convergence performance and better distribution than NSGA-Ⅱ and MOEA/D.

关 键 词：大规模多目标优化 分解策略 粒子群优化算法 

分 类 号：TP30[自动化与计算机技术—计算机系统结构]