详细信息
基于拥挤度与变异的动态微粒群多目标优化算法 ( EI收录)
Improved PSO-based multi-objective optimization by crowding with mutation and particle swarm optimization dynamic changing
文献类型:期刊文献
中文题名:基于拥挤度与变异的动态微粒群多目标优化算法
英文题名:Improved PSO-based multi-objective optimization by crowding with mutation and particle swarm optimization dynamic changing
作者:王辉[1,2];钱锋[1,2]
机构:[1]华东理工大学化学工程联合国家重点实验室;[2]华东理工大学信息科学与工程学院,上海200237
年份:2008
卷号:23
期号:11
起止页码:1238
中文期刊名:控制与决策
外文期刊名:Control and Decision
收录:CSTPCD;;EI(收录号:20085111797507);Scopus;北大核心:【北大核心2004】;CSCD:【CSCD2011_2012】;
基金:国家杰出青年科学基金项目(60625302);国家973计划项目(2002CB3122000);国家863计划项目(20060104Z1081);国家自然科学基金项目(60704028);国家科技支撑计划项目(2007BAF22B05);上海市科委重大基础研究项目(05DJ14002);上海市基础研究重点项目(07JC14016)
语种:中文
中文关键词:微粒群优化;多目标优化;动态变化;拥挤度;变异操作
外文关键词:Particle swarm optimization; Muhi-ohjective optimization; Dynamic changing; Crowding; Mutation
摘要:提出一种动态微粒群多目标优化算法(DCMOPSO),算法中的惯性权重和加速因子动态变化以增强算法的全局搜索能力,并采用拥挤度的方法对外部档案进行维护以增加非劣解的多样性.在维护过程中,从外部档案中按拥挤度为每个微粒选择全局最好位置,同时使用变异操作避免算法早熟.通过几个典型的多目标测试函数对DCMOPSO算法的性能进行了测试,并与多目标优化算法MOPSO和NSGA-Ⅱ进行对比.结果表明,DCMOPSO算法具有良好的搜索性能.
A dynamic changing multi-objection particle swarm optimization (DCMOPSO) algorithm is proposed, in which inertia weight and acceleration coefficients are dynamic changing to explore the search space more efficiently. Crowding distance is used to maintain the external archive and mutation operator mechanism is also adopted to maintain the diversity of non-dominated solutions. The global best location for every particle is selected in the procedure of external archive maintenance. Some benchmark functions are tested for comparing the performance of DCMOPSO with MOPSO and NSGA- Ⅱ. The results show the feasibility of DCMOPSO.
参考文献:
正在载入数据...
