多策略协同改进蜣螂优化算法  

Multi-strategy Collaborative Improvement Dung Beetle Optimization Algorithm

作　　者：方耀楚 刘水平 卓菁嫄 陈卫 陈婉若 FANG Yaochu;LIU Shuiping;ZHUO Jingyuan;CHEN Wei;CHEN Wanruo(School of Civil Engineering,University of South China,Hengyang,Hunan 421001,China;Hunan Provincial Key Laboratory of High Performance Special Concrete,Hengyang,Hunan 421001,China;China Nuclear Construction Key Laboratory of High Performance Concrete,Hengyang,Hunan 421001,China)

机构地区：[1]南华大学土木工程学院,湖南衡阳421001 [2]高性能混凝土湖南省重点实验室,湖南衡阳421001 [3]中国核建高性能混凝土重点实验室,湖南衡阳421001

出　　处：《南华大学学报（自然科学版）》2024年第4期1-16,共16页Journal of University of South China：Science and Technology

基　　金：湖南省自然科学基金项目(S2021JJSSLH0071);湖南省教育厅优秀青年资助项目(22B0443;306247)。

摘　　要：针对蜣螂优化算法(Dung beetle optimizer,DBO)全局搜索和局部开发能力不平衡、全局搜索能力弱、易陷入局部解的缺点,提出一种多策略协同改进的蜣螂优化算法(Dung beetle optimizer of PSO,简称DBPSO)。首先,使用Piecewise混沌映射初始化种群,使初始解位置更均匀,增加种群的丰富性;其次,引入改进正余弦算法,协调全局勘探和局部开发能力;然后,通过引入非线性衰减因子调节莱维飞行和布朗运动和加入警戒蜣螂机制对蜣螂最优位置进行扰动。通过CEC2005和CEC2019测试函数和Wilcoxon秩和检验,与多种元启发式算法对比验证了DBPSO算法具有很好的性能。最后,为进一步说明DBPSO算法在实际问题中的应用潜力,将3个实际工程设计问题进行求解,实验结果表明,所提DBPSO算法对于实际工程问题能有效地求解。To address the shortcomings of the Dung Beetle Optimizer(DBO),which has an imbalance between global search and local exploitation ability,weak global search ability,and easy to fall into local solutions,a multi-strategy collaborative improved dung beetle optimization algorithm(DBPSO)is proposed.First,the population is initialized u sing piecewise chaotic mapping to enhance the uniformity of initial solution locations and increase the population diversity;second,an improved positive cosine algorithm is introduced to coordinate the global exploration and local exploitation capailities;next,the optimal location of the dung beetle is perturbed by introducing nonlinear decay factors to regulate the Lévy flights and Brownian motions and by incorporating a vigliant dung beetle mechanism.The DBPSO algorithm is verified to have good performance by CEC2005 and CEC2019 test functions and Wilcoxon rank sum test,in comparison with multiple meta-heuristic algorithms.Finally,to further illustrate the potential of the DBPSO algorithm in practical problems,three real engineering design problems are solved,and the experimental results show that the proposed DBPSO algorithm is effective for practical engineering problems.

关 键 词：蜣螂优化算法 DBPSO 混沌映射 莱维飞行 CEC2005测试函数 CEC2019测试函数 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]