Opposition-based differential evolution algorithm with Gaussian distribution estimation
摘要:
针对差分进化算法探索能力不足、收敛慢等问题,提出一种基于高斯分布估计的对位差分进化算法.该算法在生成对位种群的同时还生成一个基于高斯分布估计的新种群,意在更充分地搜索解空间.在不满足跳转条件的情况下,算法给出一种基于高斯分布估计的种群跳转,增加了种群多样性.在选择操作时,将所有父代和子代个体混合起来择优选择,减少了部分优秀解和优秀基因的流失.最后在CEC2014标准函数中进行测试,与其他算法进行比较,验证了所提出的算法具有更好的搜索能力和收敛性.
Aiming at the problems of inefficient exploration and slow convergence,an Opposition-based Differential Evolution(ODE)with Gaussian distribution estimation is proposed in this paper.The algorithm generates a population based on Gaussian distribution estimation while it generates the opposite population,which fully explores the solution space.When the jumping condition is not met in ODE,a new jumping population is generated with Gaussian distribution estimation for well population diversity.During the selection operation,all the parents and children are mixed together for the best selection,which reduces the loss of some excellent solutions and genes.Finally,based on CEC2014 benchmark,the proposed algorithm is compared with other differential evolution algorithms.Experimental results indicate that the proposed algorithm has stronger search ability and better convergence property.
作者:
方景远 季益胜 赵新超
Fang Jingyuan;Ji Yisheng;Zhao Xinchao(School of Science,Beijing University of Posts and Telecommunications,Beijing 100876,China)
机构地区:
北京邮电大学理学院
出处:
《betway官方app 学报:自然科学版》 CAS 北大核心 2021年第3期27-32,共6页
基金:
国家自然科学基金(61973042) 北京市自然科学基金(1202020) 北京邮电大学提升科技创新能力行动计划项目(2020XD-A01-1).
关键词:
差分进化算法 对位学习 高斯分布估计
differential evolution algorithm opposition-based learning Gaussian distribution estimation
分类号:
TP278 [自动化与计算机技术—检测技术与自动化装置] TP391 [自动化与计算机技术—计算机应用技术]