基于聚类分析的差分算法协作研究
|
摘要
针对差分进化算法存在的收敛速度慢、易陷入局部最优等不足,本文提出一种融入聚类分析的差分进化算法。首先,利用聚类分析方法将差分算法的种群进行聚类分类,抽取代表元个体,利用新的个体来替换原种群中的较差个体,去除种群中的冗余信息将种群进行优化更新,从而使得整个种群可以快速准确地收敛于全局最优解。最后本文利用MATLAB编程模拟仿真,基于CEC2005测试函数库进行了模拟实验,结果表明加入了聚类分析替换策略的差分进化算法不仅有效地抑制了早熟收敛、提高了收敛速度,还有着简洁高效、鲁棒性强等特性。
Aiming at the shortcomings of differential evolution(DE) algorithm,such as slow convergence speed and easy to fall into local optimal solution,a differential evolution algorithm that incorporates cluster analysis is proposed in this paper.Cluster analysis is firstly used to classify populations,and typical new individuals are extracted.New individuals are used to replace poor individuals in the original population and redundant information in the population is removed.The population is optimized and updated so that the entire population can quickly and accurately converge to the global optimum.Finally,this paper does simulation experiments with CEC2005 test function using MATLAB simulation.The results show that differential evolution algorithm with cluster analysis replacing strategy not only effectively inhibits premature convergence,improves the convergence speed,but also has simple,efficient,and robust characteristics.
Aiming at the shortcomings of differential evolution(DE) algorithm,such as slow convergence speed and easy to fall into local optimal solution,a differential evolution algorithm that incorporates cluster analysis is proposed in this paper.Cluster analysis is firstly used to classify populations,and typical new individuals are extracted.New individuals are used to replace poor individuals in the original population and redundant information in the population is removed.The population is optimized and updated so that the entire population can quickly and accurately converge to the global optimum.Finally,this paper does simulation experiments with CEC2005 test function using MATLAB simulation.The results show that differential evolution algorithm with cluster analysis replacing strategy not only effectively inhibits premature convergence,improves the convergence speed,but also has simple,efficient,and robust characteristics.
引文
[1]Beni G,Wang J.,Swarm Intelligence in Cellular Robotic Systems[M].Robots and Biological Systems:Towards a New Bionics.1993:703-712.
[2]Holland JH,Adaptation in natural and artificial systems[J].Ann Arbor,1975,6(2):126-137.
[3]Storn and K.Price,Differential evolution a simple and efficient heuristic for global optimization over continuous spaces[J].Journal of Global Optimization,vol.11,no.4,pp.341-359,1997.
[4]Pe?u?uri F,Cab C,Carvente O,et al.A Study of the Classical Differential Evolution Control Parameters[J].Swarm&Evolutionary Computation,2016,26:86-96.
[5]吕铭晟,沈洪远,李志高,等.多变异策略差分进化算法的研究与应用[J].计算机工程,2014,40(12):146-150.
[6]Cui C Y,Jiao Y C,Zhang L.Synthesis of Some Low Sidelobe Linear Arrays Using Hybrid Differential Evolution Algorithm Integrated with Convex Programming[J].IEEEAntennas&Wireless Propagation Letters,2017,16(99):2444-2448.
[7]Das S,Abraham A,Konar A.Particle Swarm Optimization and Differential Evolution Algorithms:Technical Analysis,Applications and Hybridization Perspectives[J].2010,116:1-38.
[8]Wang P C,Qian X,Zhou Y,et al.A novel Differential Evolution algorithm based on simulated annealing[C]//Chinese Control and Decision Conference.20-10:7-10.
[9]张英杰,龚中汉.基于阈值统计学习的差分进化引力搜索算法[J].计算机研究与发展,2014,51(10):2187-2194.
[10]Brest J,Greiner S,Boskovic B,et al.Self-Adapting Control Parameters in Differential Evolution:A Comparative Study on Numerical Benchmark Problems[J].IEEE Transactions on Evolutionary Computation,2006,10(6):646-657.2006.
[11]Pang-NingTan,MichaelSteinbach,VipinKumar.数据挖掘导论:完整版.第2版[M].人民邮电出版社,2011.
[12]左黎斌,何傲,王昕,等.基于FCM聚类算法的电能表标准装置监测数据分析与研究[J].软件,2018,39(6):89-95.
[13]何傲,左黎斌,王昕,等.基于K-means算法的电能表检定误差分析与研究[J].软件,2018,39(6):64-73.
[14]陈曦斌,焦明海,刘昊汧,等.移动机器人养老服务路径规划的粒子群算法研究[J].软件,2018,39(6):135-138.
[15]刘露萍,贾文生.基于方体剖分和量子免疫粒子群算法的Nash均衡求解[J].软件,2018,39(6):01-03.
[16]颜乐鸣.基于工作流的软件测试过程模型研究[J].软件,2018,39(5):160-165.
[2]Holland JH,Adaptation in natural and artificial systems[J].Ann Arbor,1975,6(2):126-137.
[3]Storn and K.Price,Differential evolution a simple and efficient heuristic for global optimization over continuous spaces[J].Journal of Global Optimization,vol.11,no.4,pp.341-359,1997.
[4]Pe?u?uri F,Cab C,Carvente O,et al.A Study of the Classical Differential Evolution Control Parameters[J].Swarm&Evolutionary Computation,2016,26:86-96.
[5]吕铭晟,沈洪远,李志高,等.多变异策略差分进化算法的研究与应用[J].计算机工程,2014,40(12):146-150.
[6]Cui C Y,Jiao Y C,Zhang L.Synthesis of Some Low Sidelobe Linear Arrays Using Hybrid Differential Evolution Algorithm Integrated with Convex Programming[J].IEEEAntennas&Wireless Propagation Letters,2017,16(99):2444-2448.
[7]Das S,Abraham A,Konar A.Particle Swarm Optimization and Differential Evolution Algorithms:Technical Analysis,Applications and Hybridization Perspectives[J].2010,116:1-38.
[8]Wang P C,Qian X,Zhou Y,et al.A novel Differential Evolution algorithm based on simulated annealing[C]//Chinese Control and Decision Conference.20-10:7-10.
[9]张英杰,龚中汉.基于阈值统计学习的差分进化引力搜索算法[J].计算机研究与发展,2014,51(10):2187-2194.
[10]Brest J,Greiner S,Boskovic B,et al.Self-Adapting Control Parameters in Differential Evolution:A Comparative Study on Numerical Benchmark Problems[J].IEEE Transactions on Evolutionary Computation,2006,10(6):646-657.2006.
[11]Pang-NingTan,MichaelSteinbach,VipinKumar.数据挖掘导论:完整版.第2版[M].人民邮电出版社,2011.
[12]左黎斌,何傲,王昕,等.基于FCM聚类算法的电能表标准装置监测数据分析与研究[J].软件,2018,39(6):89-95.
[13]何傲,左黎斌,王昕,等.基于K-means算法的电能表检定误差分析与研究[J].软件,2018,39(6):64-73.
[14]陈曦斌,焦明海,刘昊汧,等.移动机器人养老服务路径规划的粒子群算法研究[J].软件,2018,39(6):135-138.
[15]刘露萍,贾文生.基于方体剖分和量子免疫粒子群算法的Nash均衡求解[J].软件,2018,39(6):01-03.
[16]颜乐鸣.基于工作流的软件测试过程模型研究[J].软件,2018,39(5):160-165.