摘要
差分进化(differential evolution,DE)算法是一种种群随机搜索算法,但其在收敛过程中存在着容易陷入局部最优、收敛精度不高等问题.为更好地提升DE算法的性能,提出一种改进算法为基于反向学习和伯恩斯坦算子的差分进化算法.反向学习策略用于增加种群的多样性,扩大种群的搜索范围,从而弥补陷入局部最优的不足,提高了收敛速度;伯恩斯坦多项式随机产生算法的结构参数值控制了进化过程中的突变和交叉阶段,改变了差分进化算法原有的进化策略,提高了收敛性能,是一种更加快速、高效的无参数方法.通过国际标准测试函数的实验结果表明,改进后的差分进化算法具有更强的全局寻优能力,整体收敛速度和精度明显提高.
Differential evolution(DE)algorithm is a population random search algorithm.However,in the process of convergence,it is easy to fall into local optimal value and the convergence accuracy is not high.In order to improve the performance of DE algorithm,a differential evolution algorithm based on opposition-based learning and Bernstain operator(BODE)is proposed in this article.Opposition-based learning is used to increase the diversity of the population,expand the search range of the population,so as to alleviate falling into local optimal value and improve the convergence speed.Bernstain polynomial randomly generates the structural parameter values of the algorithm,which controls the mutation and crossover stages in the evolution process,changes the original evolution strategy of the differential evolution algorithm,improves the convergence performance,and is a more rapid and efficient method without parameters.The experimental results of the international standard test function show that the improved differential evolution algorithm has stronger global optimization ability.The overall convergence speed and accuracy have significantly improved.
作者
熊聪聪
杨晓艺
王丹
李俊伟
管祥烁
XIONG Congcong;YANG Xiaoyi;WANG Dan;LI Junwei;GUAN Xiangshuo(College of Artificial Intelligence,Tianjin University of Science&Technology,Tianjin 300457,China)
出处
《天津科技大学学报》
CAS
2022年第1期46-51,共6页
Journal of Tianjin University of Science & Technology
基金
国家自然科学基金资助项目(11803022)。
关键词
差分进化算法
伯恩斯坦多项式
反向学习
differential evolution algorithm
Bernstain polynomial
opposition-based learning
