摘要
针对传统的粒子群优化算法搜索孤立区域效果差、搜索精度低等缺点,提出一种子群分层的粗粒度粒子群优化算法。在粗粒度模型的基础上,将子群分为若干普通子群、自适应子群和精英子群,不同的子群在进化过程中采取不同的进化策略。普通子群根据种群的早熟收敛程度和粒子的适应度值自适应调整惯性权重,自适应子群的速度和位置更新受到普通子群中的全局最优个体影响,精英子群保存普通子群和自适应子群的全局最优个体,并采用免疫克隆机制保证其多样性。仿真结果表明了所提算法的优异性。
For the disadvantages of traditional particle swarm optimization(PSO)in searching for isolated regions and its poor search accuracy,a coarse-grained particle swarm optimization model based on subgroup stratification was proposed.On the basis of the coarse-grained model,the subgroup was divided into several ordinary subgroups,an adaptive subgroup and an elite subgroup,and different subgroups adopted different evolutionary strategies in the evolutionary process.The ordinary subgroups adjusted inertia weight according to premature convergence degree and the fitness of the particle,the adaptive subgroup's speed and position update were adjusted by global optimal individual subgroups in ordinary subgroups,elite subgroups stored the global optimal particles owned by ordinary subgroups and the adaptive subgroup and the immune clone mechanism were used to ensure its diversity,the simulation result indicates the superiority of the optimization.
引文
[1]ZHU Degang,SUN Hui,ZHAO Jia,et al.Particle swarm optimization algorithm based on Gaussian perturbation[J].Journal of Computer Applications,2014,34(3):754-759(in Chinese).[朱德刚,孙辉,赵嘉,等.基于高斯扰动的粒子群优化算法[J].计算机应用,2014,34(3):754-759.]
[2]Kyle Robert Harrison,Andries P,Engelbrecht,et al.Inertia weight control strategies for particle swarm optimization[J].Swarm Intelligence,2016,10(4):267-305.
[3]LI Zhe,BAI Peng,TAN Ruilian,et al.Research on acceleration factor adjustment strategy in particle swarm optimization[C]//International Conference on Computer,Communications and Information Technology,2014:380-383(in Chinese).[李哲,柏鹏,檀蕊莲,等.粒子群算法中加速因子调整策略研究[C]//计算机,通信和信息技术国际会议,2014:380-383.]
[4]FAN Chengli,XING Qinghua,FAN Haixiong,et al.Variable neighborhood particle swarm optimization algorithm with scarcity factor[J].Control and Decision,2014,29(4):696-700(in Chinese).[范成礼,邢清华,范海雄,等.带审敛因子的变邻域粒子群算法[J].控制与决策,2014,29(4):696-700.]
[5]Li Kewen,Shao Mingwen, Wu Weizhi. A data reduction method in formal fuzzy contexts[J].International Journal of Machine Learning and Cybernetics,2017,8(4):1145-1155.
[6]Li Kewen,Liu Lu,Zhai Jiannan,et al.The improved grey model based on particle swarm optimization algorithm for time series prediction[J].Engineering Applications of Artificial Intelligence,2016,55(C):285-291.
[7]JIN Min,LU Huaxiang.A hierarchical hybrid algorithm for multi-subgroups based on genetic algorithm and particle swarm optimization[J].Control Theory&Applications,2013,30(10):1231-1238(in Chinese).[金敏,鲁华祥.一种遗传算法与粒子群优化的多子群分层混合算法[J].控制理论与应用,2013,30(10):1231-1238.]
[8]TANG Yiling,JIANG Shunliang,YE Famao,et al.Particle swarm optimization algorithm for optimal particle enhancement[J].Computer Engineering and Applications,2017,53(4):25-32(in Chinese).[唐祎玲,江顺亮,叶发茂,等.最优粒子增强探索粒子群算法[J].计算机工程与应用,2017,53(4):25-32.]
[9]Garvishkumar K Patel,Vipul K Dabhi,Harshadkumar B,et al.Clustering using a combination of particle swarm optimization and K-means[J].Journal of Intelligent Systems,2017,26(3):457-469.
[10]HE Cheng,LI Meiyi,QIU Qianqian.Differential information evolutionary particle swarm optimization algorithm[J].Application Research of Computers,2012,29(6):2047-2051(in Chinese).[何诚,李枚毅,邱茜茜.全信息差异进化粒子群优化算法[J].计算机应用研究,2012,29(6):2047-2051.]
[11]Wei Hong Lim,Nor Ashidi Mat Isa.An adaptive two-layer particle swarm optimization with elitist learning strategy[J].Information Sciences,2014,273(3):49-72.