摘要
针对标准灰狼优化算法(GWO)易陷入局部最优和求解精度低的问题,提出一种基于Logistic模型的控制参数自适应调整GWO(AGWO)算法.分析了控制参数a在算法进化过程中的重要作用,将Logistic模型理论嵌入到GWO算法中自适应调整控制参数a.此外,为了提高算法的全局收敛速度,用混沌序列方法产生初始种群.采用8个复杂基准测试函数进行数值实验,在相同的最大适应度函数评价次数下,AGWO总体性能上均优于标准GWO、NGWO、GWO-DE、IGWO和GA-GWO算法.实验结果表明,在GWO算法框架内,采用Logistic模型自适应调整控制参数在性能上明显优于线性递减调整方式.
Aimed at the problem that the solution of the grey wolf optimization(GWO)algorithm is ease to be trapped into local optimum and has low precision,an adaptive grey wolf optimization GWO(AGWO)algorithm with Logistic model-base control parameter was proposed.The important role of the control parameter a in the evolution process of the algorithm was analyzed.The Logistic model theory was imbedded into the GWO algorithm to make adaptive adjustment of the control parameter a.In addition,in order to enhance the global convergence speed of the algorithm an initial population was generated with chaotic sequence method.It was shown by numeric experiment on 8 complex standard test functions that the overall performance of AGWO was superior to that of the standard GWO,NGWO,GWO-DE,IGWO,and GA-GWO when the number of evaluation of maximum fitness function was identical.The simulation result showed that in framework of GWO algorithm the adaptive adjustment of control parameter with Logistic model would remarkably be superior to the linear progressive decrease adjustment method in connection with its performance.
作者
陈清容
唐斌
CHEN Qing-rong;TANG Bin(School of Information,Guizhou University of Finance and Economics,Guiyang 550025,China)
出处
《兰州理工大学学报》
CAS
北大核心
2018年第2期95-101,共7页
Journal of Lanzhou University of Technology
基金
贵州省科学技术基金(黔科合基础[2016]1022)
