基于局部协同与竞争变异的动态多种群粒子群算法
投稿时间:2020-12-11  修订日期:2021-01-14  点此下载全文
引用本文:
摘要点击次数: 104
全文下载次数: 0
作者单位邮编
孙欣 上海工程技术大学 电子电气工程学院 201620
于慧 中国移动通信集团上海有限公司闵行分公司 
王宇嘉* 上海工程技术大学 电子电气工程学院 201620
林炜星 上海工程技术大学 电子电气工程学院 
梁海娜 上海工程技术大学 电子电气工程学院 
陈万芬 上海工程技术大学 电子电气工程学院 
基金项目:国家自然科学基金资助项目(61403249);上海工程技术大学研究生创新项目(19KY0210)
中文摘要:为了解决多目标优化问题中多样性与收敛性之间的不平衡问题,取得更好的优化结果,提出了基于局部协同与竞争变异的动态多种群粒子群算法。该算法利用局部协同将初始种群进行划分,形成的子种群分别采用非支配排序方法和差分变异方法选择各自种群的领导粒子。为了提高算法的局部搜索能力,改进粒子的更新方式,使得各个目标优化更加均衡。为了防止出现“早熟”收敛的情况,引入竞争变异来增加种群多样性。最后,通过选择一系列标准测试函数将该算法与现存算法进行比较实验,验证该算法在多目标优化问题上的有效性。实验结果显示,该算法的多样性和收敛性比其它算法更好,优化效果更佳。
中文关键词:多目标优化问题  粒子群算法  多种群  局部协同  竞争变异
 
Dynamic multi-population particle swarm optimization based on local cooperative and competitive variation
Abstract:In order to solve the imbalance between diversity and convergence in multi-objective optimization problem and obtain better optimization results, a dynamic multi-population particle swarm optimization algorithm based on local cooperative and competitive variation is proposed. The initial population is divided by cooperative strategy, and the resulting subpopulation is selected by non-dominant sorting method and differential variation method respectively in this algorithm. Improving the ability of the algorithm’s local searching and changing the updating method of particles, the optimization of each target is more balanced. In order to prevent premature convergence, competitive mutation strategy is introduced to promote population diversity. Finally, by selecting a series of standard test functions, the algorithm is compared with existing algorithms to verify its effectiveness in multi-objective optimization. Experimental results show that the algorithm has better diversity and convergence than other algorithms, and the optimization effect is better.
keywords:Multi-objective optimization problem (MOP)  particle swarm optimization (PSO)  multi-population  local cooperative  Competition mutation
查看全文   查看/发表评论   下载pdf阅读器