个体扰动的混沌对立学习与差分进化灰狼算法  

Grey wolf optimization algorithm integrating individual disruption based on chaotic opposition-learning and differential evolution

作　　者：崔建弘[1] 林海霞[1] 吕晓华[1] 张卫娟[1] CUI Jian-hong;LIN Hai-xia;LYU Xiao-hua;ZHANG Wei-juan(College of Artificial Intelligence and Big Data,Hebei Polytechnic Institute,Shijiazhuang 050000,China)

机构地区：[1]河北工程技术学院人工智能与大数据学院,河北石家庄050000

出　　处：《计算机工程与设计》2022年第2期587-595,共9页Computer Engineering and Design

基　　金：国家自然科学基金项目(51305152);河北工程技术学院校级基金项目(2020HG07);新世纪优秀人才支持计划基金项目(NCET-07-0745);河北省教育科学“十三五”规划课题基金项目(1904345);河北省科技支撑计划基金项目(16210804)。

摘　　要：针对传统灰狼优化算法易于陷入局部最优、寻优精度低的问题,提出基于混沌对立学习和差分进化机制的改进灰狼优化算法CODEGWO。引入混沌对立学习策略生成灰狼初始种群,提升初始解的质量,加速算法收敛;引入差分进化的局部搜索机制,改善灰狼的局部开发与邻近区域的搜索能力;引入个体扰动机制增加种群多样性,改进灰狼的全局搜索能力。8个单峰和多峰基准函数优化求解的测试结果表明,CODEGWO算法可以有效提升寻优精度和收敛速度。The traditional gray wolf optimization algorithm is easy to fall into local optimum and low accuracy.For solving this problem,an improved grey wolf optimization algorithm CODEGWO based on chaotic opposition-learning and differential evolution mechanism was presented.A chaotic opposition-learning strategy was introduced to generate the initial grey population,which promoted the quality of the initial solutions and accelerated the convergence of the algorithm.The differential evolution was brought in the local search mechanism for improving the local exploitation of greys and the search ability for adjacent area.The individual disruption mechanism was introduced for augmenting the population diversity and ameliorating the global search ability of greys.The test results of eight benchmark functions optimization solving consisting of unimodal and multimodal functions show that,CODEGWO can effectively improve optimization precision and convergence speed.

关 键 词：灰狼优化算法 对立学习 混沌系统 差分进化 个体扰动 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]