计算机工程与应用 ›› 2021, Vol. 57 ›› Issue (22): 92-101.DOI: 10.3778/j.issn.1002-8331.2011-0409

• 理论与研发 • 上一篇    下一篇

分段权重和变异反向学习的蝴蝶优化算法

李守玉,何庆,杜逆索   

  1. 1.贵州大学 大数据与信息工程学院,贵阳 550025
    2.贵州大学 贵州省大数据产业发展应用研究院,贵阳 550025
  • 出版日期:2021-11-15 发布日期:2021-11-16

Piecewise Weight and Mutation Opposition-Based Learning Butterfly Optimization Algorithm

LI Shouyu, HE Qing, DU Nisuo   

  1. 1.College of Big Data & Information Engineering, Guizhou University, Guiyang 550025, China
    2.Guizhou Big Data Academy, Guizhou University, Guiyang 550025, China
  • Online:2021-11-15 Published:2021-11-16

摘要:

针对原始蝴蝶优化算法容易陷入局部最优解、收敛速度慢及寻优精度低等问题,提出分段权重和变异反向学习的蝴蝶优化算法。通过飞行引领策略来矫正邻域内蝴蝶的自身飞行,降低盲目飞行,增强算法跳出局部最优的能力;引入分段权重来平衡全局勘探及局部开发的能力,进而实现蝴蝶位置动态更新;使用变异反向学习对位置进行扰动,增加种群多样性以及提高算法的收敛速度。通过对9个测试函数和部分CEC2014函数及Wilcoxon秩和检验来评估改进算法的寻优能力,实验结果表明改进算法的收敛速度及寻优精度得到了极大改进。

关键词: 蝴蝶优化算法(BOA), 飞行引领策略, 分段权重, 变异反向学习, 统计检验

Abstract:

In order to solve the problems of the original butterfly optimization algorithm, such as local optimal solution, slow convergence speed and low searching precision, a piecewise weight and mutation opposition-based learning butterfly optimization algorithm is proposed. Firstly, the flight guidance strategy is adopted to correct the flight of butterflies in the neighborhood, reduce blind flight, and enhance the ability of the algorithm to jump out of local optimal solution. Then, the segmented weight is introduced to balance the ability of global exploration and local development, so as to realize the dynamic updating of butterfly position. Finally, the mutation opposition-based learning is used to perturb the location, increase the population diversity and improve the convergence rate of the algorithm. The optimization ability of the improved algorithm is evaluated on 9 test functions, partial CEC2014 functions and Wilcoxon rank sum test. The experimental results show that the convergence speed and optimization accuracy of the improved algorithm are greatly improved.

Key words: Butterfly Optimization Algorithm(BOA), flight guidance strategy, piecewise weight, mutation opposition-based learning, statistical test