一种基于相角映射的改进多目标粒子群优化算法  

Improved multi-objective particle swarm optimization algorithm based on phase angle reflection

作　　者：李婷[1,2] 吴敏[1,2] 何勇[1,2] 

机构地区：[1]中南大学信息科学与工程学院,长沙410083 [2]中南大学先进控制与智能自动化湖南省工程实验室,长沙410083

出　　处：《控制与决策》2013年第10期1513-1519,共7页Control and Decision

基　　金：国家自然科学基金重大国际合作研究项目(61210011);国家自然科学基金项目(60974045)

摘　　要：提出一种相角粒子群优化算法求解多目标优化问题.该算法采用相角映射实现了粒子在相角空间上仅依赖于归一化多目标函数的快速搜索,在粒子飞行信息共享机制上引入共享池概念,提出基于关联支配排序和相似度排序的共享池更新策略,提高了Pareto解的多样性.采用Sigma领导策略和混沌变异操作,平衡了算法的快速搜索能力和全局寻优能力.标准多目标测试函数和电力系统广域阻尼控制多目标优化算例表明了所提出算法的可行性和有效性.Phase angle particle swarm optimization（PAPSO） algorithm is proposed to solve multi-objective optimization problems. The algorithm adopts phase angle reflection, such that particles can rapidly search solutions in phase angle space, which is only dependent on normalized multi-objective functions. Sharing pool concept is introduced into the particle flight information sharing mechanism, and the sharing pool update strategy based on the related predominance sorting and similarity degree sorting is presented to improve the diversity of Pareto solutions. Sigma leading strategy and chaos variation operation balance its ability of finding optimal solutions rapidly and extensively. Cases related to benchmark multi- objective test functions and wide area damping control multi-objective optimization in power systems verify the feasibility and effectiveness of the proposed algorithm.

关 键 词：多目标优化 粒子群优化算法 混沌 多目标粒子群优化算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]