面向分布式计算的混合维度微粒群算法
|
摘要
在智能计算方面,微粒群算法有着搜索速度快、算法实现简单、参数设置少等优点而得到了广泛的应用。但遇到大规模问题时会遇到容易陷入局部最优、计算耗时、结果精度不高的情况。因此本文提出以分布式计算方法为基础的混合维度微粒群算法MDPSO(Mixed Dimension Particle Swarm Optimization),该算法采用分布式计算方式实现任务并行,通过混合维度的方法进行种群间的信息交流。最后通过对5个测试函数和岛屿模型PSO算法进行实验对比,结果证明所提出的算法能够更好的跳出局部最优而得到更为准确的结果。
Due to the advantages of fast searching speed,easy implementation and few parameters requirement in area of intelligent computation,the particle swarm algorithm has gotten a wide application in evolutionary computation. However,when facing with large-scale issues,the particle swarm algorithm has limitation of local optimization,time-consuming calculation and low result accuracy. Therefore,this paper proposes a mixed dimension particle swarm optimization( MDPSO) algorithm which uses the distributed computing method to achieve the task parallelism,and exchanges information among populations by the method of mixed dimensions. Finally,the experimental results show that the proposed algorithm can jump out easily of the local optimal and get more accurate results by comparing the five test functions with the Island Model PSO algorithm.
Due to the advantages of fast searching speed,easy implementation and few parameters requirement in area of intelligent computation,the particle swarm algorithm has gotten a wide application in evolutionary computation. However,when facing with large-scale issues,the particle swarm algorithm has limitation of local optimization,time-consuming calculation and low result accuracy. Therefore,this paper proposes a mixed dimension particle swarm optimization( MDPSO) algorithm which uses the distributed computing method to achieve the task parallelism,and exchanges information among populations by the method of mixed dimensions. Finally,the experimental results show that the proposed algorithm can jump out easily of the local optimal and get more accurate results by comparing the five test functions with the Island Model PSO algorithm.
引文
[1]蒋晓屾.粒子群算法在多维优化问题中的改进研究[D].杭州:浙江理工大学,2016:17-25.
[2]黄芳,樊晓平.基于岛屿群体模型的并行粒子群优化算法[J].控制与决策,2006,21(2):175-179.
[3] D'ANGELO G. Parallel and distributed simulation from many cores to the public cloud[C]//International Conference on High Performance Computing and Simulation,2011:14-23.
[4] BAILEY D H,Le FTON L. Review of An Introduction to Parallel and Vector Scientific Computing[M]. Cambridge uniersity press,New York,2006.
[5] SUNG S,YOUN C,KONG E,et al. A distributed mobile cloud computing model for secure big data[C]//International Conference on Information Networking,Irvine CA,USA,2016.
[6] ZHAO J,WANG L,JIE T,et al. A security framework in G-Hadoop for big data computing across distributed Cloud data centres[J]. Journal of Computer&System Sciences,2014,80(5):994-1007.
[7]李勇,吴敏,曹卫华,等.基于粒度分布评估与优化的制粒过程PSO-BP控制算法[J].自动化学报,2012,38(6):1007-1016.
[8]李志坚,吴晓军,任哲坡,等.基于分布式粗粒度并行计算的遗传规划算法研究[J].计算机应用研究,2015,32(1):48-50.
[9]唐俊. PSO算法原理及应用[J].计算机技术与发展,2010,20(2):213-216.
[10]张艳.分布并行算法设计、分析与实现[D].成都:电子科技大学,2001:27-37.
[11] YAO X,LIU Y,LIN G. Evolutionary programming made faster[J]. Evolutionary Computation IEEE Transactions,1999,3(2):82-102.
[12]程磊生,吴志健,彭虎,等.基于RDD的分布式粒子群优化算法[J].小型微型计算机系统,2016,37(11):2542-2546.
[2]黄芳,樊晓平.基于岛屿群体模型的并行粒子群优化算法[J].控制与决策,2006,21(2):175-179.
[3] D'ANGELO G. Parallel and distributed simulation from many cores to the public cloud[C]//International Conference on High Performance Computing and Simulation,2011:14-23.
[4] BAILEY D H,Le FTON L. Review of An Introduction to Parallel and Vector Scientific Computing[M]. Cambridge uniersity press,New York,2006.
[5] SUNG S,YOUN C,KONG E,et al. A distributed mobile cloud computing model for secure big data[C]//International Conference on Information Networking,Irvine CA,USA,2016.
[6] ZHAO J,WANG L,JIE T,et al. A security framework in G-Hadoop for big data computing across distributed Cloud data centres[J]. Journal of Computer&System Sciences,2014,80(5):994-1007.
[7]李勇,吴敏,曹卫华,等.基于粒度分布评估与优化的制粒过程PSO-BP控制算法[J].自动化学报,2012,38(6):1007-1016.
[8]李志坚,吴晓军,任哲坡,等.基于分布式粗粒度并行计算的遗传规划算法研究[J].计算机应用研究,2015,32(1):48-50.
[9]唐俊. PSO算法原理及应用[J].计算机技术与发展,2010,20(2):213-216.
[10]张艳.分布并行算法设计、分析与实现[D].成都:电子科技大学,2001:27-37.
[11] YAO X,LIU Y,LIN G. Evolutionary programming made faster[J]. Evolutionary Computation IEEE Transactions,1999,3(2):82-102.
[12]程磊生,吴志健,彭虎,等.基于RDD的分布式粒子群优化算法[J].小型微型计算机系统,2016,37(11):2542-2546.