一种改进的粒子群优化算法及其算法测试
|
摘要
粒子群算法原理简单、参数少、易于实现,但有时容易陷入局部最优解,收敛速度慢.本文在粒子群算法理论研究的基础上,对算法的初始值选取、惯性权重取值、算法结构进行了改进:首先采用线性惯性递减权重调整,平衡全局搜索和局部搜索的能力;然后通过logistic映射将混沌状态引入到优化变量中,增强搜索空间的遍历性;最后引入遗传算法中的选择、交叉、变异保持了种群的多样性,使其具有不易陷入局部最优的能力.采用六种典型的测试函数,对惯性权重和算法进行了测试和对比分析.结果表明,算法在收敛速度和精度上都有所提高.
Particle swarm algorithm has many advantages such as simple principle, few parameters and easy to implement, but sometimes it is easy to. get into local extremum and the convergence rate is slow. On the basis of theoretical research on algorithm, the initial value selection, inertia weight value, the structure of the algorithm are improved. Firstly,linear inertia reduction weights are adjusted to balance the global search and local search capabilities;Then chaotic states are introduced into the optimal variables through logistic mapping to enhance the ergodicity of the search space; Finally, the selection, crossover, and mutation in the genetic algorithm are introduced to maintain the diversity of the population makes it difficult to fall into the local optimum.; The genetic algorithm are introduced to maintain the diversity of the population to enhance the capability of not falling into local optimum. Six typical test functions are used to test the algorithm. The results show that the algorithm improves the convergence speed and accuracy.
Particle swarm algorithm has many advantages such as simple principle, few parameters and easy to implement, but sometimes it is easy to. get into local extremum and the convergence rate is slow. On the basis of theoretical research on algorithm, the initial value selection, inertia weight value, the structure of the algorithm are improved. Firstly,linear inertia reduction weights are adjusted to balance the global search and local search capabilities;Then chaotic states are introduced into the optimal variables through logistic mapping to enhance the ergodicity of the search space; Finally, the selection, crossover, and mutation in the genetic algorithm are introduced to maintain the diversity of the population makes it difficult to fall into the local optimum.; The genetic algorithm are introduced to maintain the diversity of the population to enhance the capability of not falling into local optimum. Six typical test functions are used to test the algorithm. The results show that the algorithm improves the convergence speed and accuracy.
引文
[1]林诗洁,董晨,陈明志,张凡,陈景辉.新型群智能优化算法综述[J].计算机工程与应用,2018(12):1-9.
[2]宋玉坚,叶春明,黄佐珲.多资源均衡优化的布谷鸟算法[J].计算机应用,2014(01):189-193.
[3]Goudarzi F, Asgari H, AI-Raweshidy H S. Traffic-Aware VANET Routing for City Environments-A Protocol Based on Ant Colony Optimization[J]. IEEE Systems Journal, 2018(99):1-11.
[4] Kenendy J, Eberh art R C. Particle Swaxrm Optimization[C]//International Conference on Neural Networks, 1995:1942-1948.
[5]王智昊,郑向伟,马红伟.一种基于改进粒子群优化和模拟退火的Memetic算法[J].小型微型计算机系统,2013, 34(1):617-620.
[6]陈君兰,叶春明.粒子群优化算法在柔性资源受限项目调度中的研究[J].计算机科学,2013, 2(02):241-244.
[7]程军,李荣钧.基于粒子群优化的神经网络预测模型[J].数学的实践与认识,2015, 45(1):176-180.
[8]石松,陈云.层次环形拓扑结构的动态粒子群算法[J].计算机工程与应用,2013, 49(8):1-6.
[9] Ma Hai, Wang Y J, Wei M A, Hu R. Application of RBF neural network in joint inversion of seismic logging based on PSO optimization[J]. Journal of Oil Gas Technology, 2008, 30(1):71-84.
[10] Da Y, Xiurun G. An improved PSO-based ANN with simulated annealing technique[J]. Neurocomputing, 2005, 63:527-533.
[11]田思琪,郎百和,韩太林.基于混沌-量子粒子群的分簇路由算法[J].吉林大学学报(信息科学版),2018,36(1):14-19.
[12]康朝海,李鹏娜,张永丰,陈建玲.基于混合蛙跳粒子群算法的TSP问题求解[J].吉林大学学报(信息科学版),2017, 35(3):498-506.
[13] Li Duan, Zhang Hongxin, Muhammad Saad Khan, Mi Fang. Recognition of motor imagery tasks for BCI using CSP and chaotic PSO twin SVM[J]. The Journal of China Universities of Posts and Telecommunications,2017, 24(1):83-90.
[14] Shi Y, Eberhart R C. Fuzzy adaptive particle swarm optimization[J].Congress on Evolutionary Computation,1999,1(12):101-106.
[15] Tang Y, Wang Z, Fang J A. Feedback learning particle swarm optimization[J]. Applied Soft Computing, 2011, 11(8):4713-4725.
[16]李会荣,高岳林,李济民.一种非线性递减惯性权重策略的粒子群优化算法[J].商洛学院学报,2007,21(2):16-20.
[17]张贝克,谢元平,马昕.结合遗传算子的改进粒子群算法在控制系统设计中的应用[J].化工自动化及仪表,2011, 38(3):521-524.
[2]宋玉坚,叶春明,黄佐珲.多资源均衡优化的布谷鸟算法[J].计算机应用,2014(01):189-193.
[3]Goudarzi F, Asgari H, AI-Raweshidy H S. Traffic-Aware VANET Routing for City Environments-A Protocol Based on Ant Colony Optimization[J]. IEEE Systems Journal, 2018(99):1-11.
[4] Kenendy J, Eberh art R C. Particle Swaxrm Optimization[C]//International Conference on Neural Networks, 1995:1942-1948.
[5]王智昊,郑向伟,马红伟.一种基于改进粒子群优化和模拟退火的Memetic算法[J].小型微型计算机系统,2013, 34(1):617-620.
[6]陈君兰,叶春明.粒子群优化算法在柔性资源受限项目调度中的研究[J].计算机科学,2013, 2(02):241-244.
[7]程军,李荣钧.基于粒子群优化的神经网络预测模型[J].数学的实践与认识,2015, 45(1):176-180.
[8]石松,陈云.层次环形拓扑结构的动态粒子群算法[J].计算机工程与应用,2013, 49(8):1-6.
[9] Ma Hai, Wang Y J, Wei M A, Hu R. Application of RBF neural network in joint inversion of seismic logging based on PSO optimization[J]. Journal of Oil Gas Technology, 2008, 30(1):71-84.
[10] Da Y, Xiurun G. An improved PSO-based ANN with simulated annealing technique[J]. Neurocomputing, 2005, 63:527-533.
[11]田思琪,郎百和,韩太林.基于混沌-量子粒子群的分簇路由算法[J].吉林大学学报(信息科学版),2018,36(1):14-19.
[12]康朝海,李鹏娜,张永丰,陈建玲.基于混合蛙跳粒子群算法的TSP问题求解[J].吉林大学学报(信息科学版),2017, 35(3):498-506.
[13] Li Duan, Zhang Hongxin, Muhammad Saad Khan, Mi Fang. Recognition of motor imagery tasks for BCI using CSP and chaotic PSO twin SVM[J]. The Journal of China Universities of Posts and Telecommunications,2017, 24(1):83-90.
[14] Shi Y, Eberhart R C. Fuzzy adaptive particle swarm optimization[J].Congress on Evolutionary Computation,1999,1(12):101-106.
[15] Tang Y, Wang Z, Fang J A. Feedback learning particle swarm optimization[J]. Applied Soft Computing, 2011, 11(8):4713-4725.
[16]李会荣,高岳林,李济民.一种非线性递减惯性权重策略的粒子群优化算法[J].商洛学院学报,2007,21(2):16-20.
[17]张贝克,谢元平,马昕.结合遗传算子的改进粒子群算法在控制系统设计中的应用[J].化工自动化及仪表,2011, 38(3):521-524.