基于动态多种群的自适应混合灰狼差分进化算法  

Differential evolution algorithm of adaptive hybrid grey wolf based on dynamic multi-population

作　　者：任晓莉 郭子璐 REN Xiao-li;GUO Zi-lu(Baoji Key Laboratory of Intelligent Information Processing,School of Computer Science,Baoji University of Arts and Sciences,Baoji 721016,Shaanxi,China;School of Law,Nankai University,Tianjin 300071,China)

机构地区：[1]宝鸡文理学院计算机学院,宝鸡市智能信息处理重点实验室,陕西宝鸡721016 [2]南开大学法学院,天津300071

出　　处：《宝鸡文理学院学报（自然科学版）》2023年第2期39-47,共9页Journal of Baoji University of Arts and Sciences(Natural Science Edition)

基　　金：国家自然科学基金项目(61971005);陕西省自然科学基础研究计划项目(2022JQ-689)。

摘　　要：目的解决灰狼算法初始种群生成方式导致其存在多样性差的问题。方法结合差分算法改进灰狼算法,生成基于动态多种群策略的混合灰狼差分算法,算法采用动态多种群策略,利用3种差分算法变异策略,增强初始种群解的多样性,并将其作为灰狼算法的初始种群,采用局部搜索策略增强算法的探索能力,最后采用改进的种群大小线性递减策略来降低算法的时间复杂度,提高算法的收敛速度。结果与结论对CEC2014测试集4类15个基准测试函数进行仿真实验,并与经典DE,GWO及其变体算法进行比较,验证了所提出算法的有效性,最后应用该算法求解经典工程领域焊接梁设计优化问题,证明了所提出算法的优越性。Purposes—To solve the problem of poor diversity caused by the initial population generation method of the grey wolf algorithm.Methods—The grey wolf algorithm is improved in combination with the difference algorithm to propose a hybrid grey wolf difference algorithm based on dynamic multi-population strategy.In the proposed algorithm,the variation strategy of three difference algorithms are used,with the dynamic multi-population strategy adopted,to enhance the diversity of the initial population solution,and take it as the initial population of the grey wolf algorithm.The local search strategy is exploited to enhance the exploration ability of the algorithm.Finally,the improved linear decline strategy of population size is utilized to reduce the time complexity of the algorithm and improve the convergence speed of the algorithm.Results and Conclusions—Simulation experiments are carried out on 15 benchmark test functions of 4 classes in CEC 2014 test set,and the effectiveness of the proposed algorithm is verified by comparing with the classical DE,GWO and its variant algorithms.Finally,the proposed algorithm is applied to solve the design optimization problem of welded beam in the classical engineering field,thus proving the superiority of the proposed algorithm.

关 键 词：灰狼算法 差分算法 动态多种群 工程优化 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]