摘要
针对差分进化算法在处理函数优化问题时存在的收敛速度较慢和过早收敛的问题,提出了一种动态参数调整的多策略差分进化算法.先将种群随机分为3个独立的子种群,分别采用3种不同的变异策略来避免种群陷入局部最优,并通过动态参数调整机制提高算法的收敛性能.经过一定代数的进化后,将种群中的优秀个体进行择优保留.采用CEC2005的25个标准测试函数对算法进行仿真,实验结果表明,新算法能够有效避免过早收敛,具有较好的优化性能.
In order to solve the problem of slow convergence and premature convergence in the differential evolution algorithm in the process of function optimization problems,a multi-strategy differential evolutionary algorithm for dynamic parameter adjustment(MDADE)is proposed.At the beginning of optimization,the population is randomly divided into three independent subpopulations.The algorithm adopts three different mutation strategies to ensure the diversity of the population.The convergence performance of the algorithm is improved by the parameter adaptive mechanism.After a certain algebraic evolution,the algorithm is simulated by 25 standard test functions of CEC2005.The experimental results show that the new algorithm can effectively avoid premature convergence and has better optimization performance.
作者
马永杰
朱琳
田福泽
MA Yong-jie;ZHU Lin;TIAN Fu-ze(School of Physics and Electronic Engineering,Northwest Normal University,Lanzhou 730070,Gansu,China)
出处
《西北师范大学学报（自然科学版）》
CAS
北大核心
2018年第3期40-46,共7页
Journal of Northwest Normal University(Natural Science)
基金
国家自然科学基金资助项目(41461078)
关键词
差分进化
择优保留
参数自适应
多策略
CEC2005
differential evolution
elitist reservation
parameter adaptive
multi-strategy
CEC2005
