基于学习与竞争的改进PSO算法研究
|
摘要
针对普通PSO算法收敛速率慢,难以收敛到全局最优解的问题,提出了一种基于学习与竞争的改进PSO算法.该算法通过将种群内部学习和竞争的思想与PSO算法相结合,让种群中个体通过竞争和学习策略来替代原有的PSO算法迭代公式.该方法在不增加PSO算法计算复杂度的基础上,能够克服基本PSO算法的不足.最后基于动态系统的稳定性分析理论,给出了该PSO算法收敛性的证明.在7种不同的测试函数上对改进后的算法进行了实验测试.实验结果表明该改进算法比传统的PSO算法有着更好的搜索精度.结果证明,新算法比普通的PSO算法具有更高的搜索精度和较低的时间复杂度.改进算法求解函数优化问题更加有效,收敛速率更快.
As PSO algorithm for the general rate of convergence is slow and difficult to converge to the global optimal solution of the problem, a competition-based learning and improvement of PSO has been proposed in this paper. The algorithm by studying populations and competitive internal thoughts and PSO algorithm combined population of individuals so that by competing with each other and learning strategies to replace the original PSO algorithm iteration formula. This method based on PSO algorithm without increasing the computational complexity, can overcome the lack of basic PSO algorithm. Enhance searching precision. On seven different test functions for improved algorithm experimentally tested. Experimental results show that the improved algorithm than the traditional PSO algorithm has better search accuracy. The results prove that the new algorithm has higher precision and lower search time complexity than conventional PSO algorithm. The improved algorithm for solving function optimization problems more efficiently, faster convergence rate.
As PSO algorithm for the general rate of convergence is slow and difficult to converge to the global optimal solution of the problem, a competition-based learning and improvement of PSO has been proposed in this paper. The algorithm by studying populations and competitive internal thoughts and PSO algorithm combined population of individuals so that by competing with each other and learning strategies to replace the original PSO algorithm iteration formula. This method based on PSO algorithm without increasing the computational complexity, can overcome the lack of basic PSO algorithm. Enhance searching precision. On seven different test functions for improved algorithm experimentally tested. Experimental results show that the improved algorithm than the traditional PSO algorithm has better search accuracy. The results prove that the new algorithm has higher precision and lower search time complexity than conventional PSO algorithm. The improved algorithm for solving function optimization problems more efficiently, faster convergence rate.
引文
[1] 蔡林益.基于粒子群算法的云计算资源配置研究 [J].西南师范大学学报(自然科学版),2017,42(9):128-132.
[2] 沈海洋.基于遗传 PSO 的无线传感网络覆盖优化算法研究 [J].微电子学与计算机,2013,30(3):148-151.
[3] 庄夏.基于并行粒子群和 RL 的无人机航路规划算法设计 [J].西南师范大学学报(自然科学版),2016,41(3):31-36.
[4] 周相兵.一种基于粒子群优化的虚拟资源分配方法 [J].重庆邮电大学学报(自然科学版),2014,26(5):686-693.
[5] 黄少荣.粒子群优化算法综述 [J].计算机工程与设计,2009,30(8):1977-1980.
[6] 王铁君,邬月春.基于混沌粒子群算法的物流配送路径优化 [J].计算机工程与应用,2011,47(29):218-221.
[7] 靳其兵,张建,权玲,等.基于混合PSO-SQP算法同时实现多变量的结构和参数辨识 [J].控制与决策,2011,26(9):1373-1376,1381.
[8] 李志华,许新,黎作鹏,等.PSO-MEA混合优化算法及其收敛性分析 [J].微电子学与计算机,2017,34(6):118-122,127.
[9] 王甫,郑亚平,刘天琪.一种基于调节因子的小生境粒子群优化算法 [J].计算机工程,2014,40(8):147-151.
[10] MAHI M,BAYKAN ? K,KODAZ H.A New Hybrid Method Based on Particle Swarm Optimization,Ant Colony Optimization and 3-Opt Algorithms for Traveling Salesman Problem [J].Applied Soft Computing,2015,30:484-490.
[11] GONG M G,CAI Q,CHEN X W,et al.Complex Network Clustering by Multiobjective Discrete Particle Swarm Optimization Based on Decomposition [J].IEEE Transactions on Evolutionary Computation,2014,18(1):82-97.
[12] 阚超豪.多向学习自适应的粒子群算法 [J].计算机工程与应用,2013,49(6):23-28.
[13] 胡勇.用随机模式和调整机制改进粒子群优化算法 [J].重庆邮电大学学报(自然科学版),2010,22(1):99-102.
[14] THANH T P,THE L N,ELNAFFAR S,et al.LPSO:Another Algorithm for Workflow Scheduling in the Cloud [J].Journal of Computer Science,2016,12(12):611-617.
[15] 李玲纯,田丽,王静.免疫粒子群算法在变电站选址中的应用 [J].贵州师范大学学报(自然科学版),2013,31(5):107-111.
[16] 张毅,代恩灿,罗元.基于改进遗传算法的移动机器人路径规划 [J].计算机测量与控制,2016,24(1):313-316.
[2] 沈海洋.基于遗传 PSO 的无线传感网络覆盖优化算法研究 [J].微电子学与计算机,2013,30(3):148-151.
[3] 庄夏.基于并行粒子群和 RL 的无人机航路规划算法设计 [J].西南师范大学学报(自然科学版),2016,41(3):31-36.
[4] 周相兵.一种基于粒子群优化的虚拟资源分配方法 [J].重庆邮电大学学报(自然科学版),2014,26(5):686-693.
[5] 黄少荣.粒子群优化算法综述 [J].计算机工程与设计,2009,30(8):1977-1980.
[6] 王铁君,邬月春.基于混沌粒子群算法的物流配送路径优化 [J].计算机工程与应用,2011,47(29):218-221.
[7] 靳其兵,张建,权玲,等.基于混合PSO-SQP算法同时实现多变量的结构和参数辨识 [J].控制与决策,2011,26(9):1373-1376,1381.
[8] 李志华,许新,黎作鹏,等.PSO-MEA混合优化算法及其收敛性分析 [J].微电子学与计算机,2017,34(6):118-122,127.
[9] 王甫,郑亚平,刘天琪.一种基于调节因子的小生境粒子群优化算法 [J].计算机工程,2014,40(8):147-151.
[10] MAHI M,BAYKAN ? K,KODAZ H.A New Hybrid Method Based on Particle Swarm Optimization,Ant Colony Optimization and 3-Opt Algorithms for Traveling Salesman Problem [J].Applied Soft Computing,2015,30:484-490.
[11] GONG M G,CAI Q,CHEN X W,et al.Complex Network Clustering by Multiobjective Discrete Particle Swarm Optimization Based on Decomposition [J].IEEE Transactions on Evolutionary Computation,2014,18(1):82-97.
[12] 阚超豪.多向学习自适应的粒子群算法 [J].计算机工程与应用,2013,49(6):23-28.
[13] 胡勇.用随机模式和调整机制改进粒子群优化算法 [J].重庆邮电大学学报(自然科学版),2010,22(1):99-102.
[14] THANH T P,THE L N,ELNAFFAR S,et al.LPSO:Another Algorithm for Workflow Scheduling in the Cloud [J].Journal of Computer Science,2016,12(12):611-617.
[15] 李玲纯,田丽,王静.免疫粒子群算法在变电站选址中的应用 [J].贵州师范大学学报(自然科学版),2013,31(5):107-111.
[16] 张毅,代恩灿,罗元.基于改进遗传算法的移动机器人路径规划 [J].计算机测量与控制,2016,24(1):313-316.