摘要
为保证在动态环境中及时跟踪到最新的真实Pareto前沿,保持解集的均匀性,提出一种基于档案交叉的动态多目标粒子群优化算法。着重利用保存在外部档案的最新非劣解,对这些非劣解进行交叉操作以增加种群的多样性,促进档案中个体信息的交流;提出一种高效的欧氏拥挤距离策略,并将其应用于对外部档案的维护;修改粒子群算法模型使之更适用于动态多目标优化。实验结果表明,该算法能适应动态环境,快速跟踪动态Pareto面,解集均匀性良好。
To track the latest true Pareto front timely in a dynamic environment and maintain the uniformity of the solution set, a dynamic multi-obj ective particle swarm optimization based on the archive crossover was presented.This algorithm emphasized on using the latest non-dominated solutions in external archives and the crossover operation was applied to these non-dominated solutions to increase the population diversity,the information exchange in the archive was promoted.An efficient Euclidean crowding distance strategy was proposed which was applied to maintain the external archive.The particle swarm optimization model was modified to make it adapt to the dynamic environment.The experimental results show that the algorithm is able to adapt to the dynamic environment and track dynamic Pareto surfaces fast and keep the solution set in good uniformity.
出处
《计算机工程与设计》
北大核心
2015年第2期507-513,共7页
Computer Engineering and Design
基金
国家自然科学基金项目(61203109)
广西空间信息与测绘重点实验室开放基金项目(桂科能1103108-16)
广西研究生教育创新计划基金项目(YCSZ2014157)
关键词
粒子群优化
动态多目标优化
外部档案
拥挤距离
交叉
particle swarm optimization
dynamic multi-obj ective optimization
external archive
crowding distance
crossover
