融合动态反向学习的阿奎拉鹰与哈里斯鹰混合优化算法

PDF下载HTML

Title:

Hybrid Aquila and Harris hawks optimization algorithm with dynamic opposition-based learning

作者:

贾鹤鸣¹, 刘庆鑫², 刘宇翔³, 王爽¹, 吴迪⁴

1. 三明学院 信息工程学院，福建 三明 365004;
2. 海南大学 计算机科学与技术学院，海南 海口 570228;
3. 福州大学 物理与信息工程学院，福建 福州 350108;
4. 三明学院 教育与音乐学院，福建 三明 365004

Author(s):

JIA Heming¹, LIU Qingxin², LIU Yuxiang³, WANG Shuang¹, WU Di⁴

1. School of Information Engineering, Sanming University, Sanming 365004, China;
2. School of Computer Science and Technology, Hainan University, Haikou 570228, China;
3. College of Physics and Information Engineering, Fuzhou University, Fuzhou 350108, China;
4. School of Education and Music, Sanming University, Sanming 365004, China

关键词:

阿奎拉鹰优化算法; 哈里斯鹰优化算法; 动态反向学习; 混合优化; 基准函数; 管柱设计问题; 汽车碰撞设计问题; Wilcoxon秩和检验

Keywords:

Aquila optimizer; Harris Hawk optimization; dynamic opposition-based learning; hybrid optimization; benchmark function; tubular column design problem; car crash design problem; Wilcoxon rank-sum test

分类号:

TP301.6

DOI:

10.11992/tis.202108031

摘要:

阿奎拉鹰优化算法(Aquila optimizer, AO)和哈里斯鹰优化算法(Harris hawks optimization, HHO)是近年提出的优化算法。AO算法全局寻优能力强，但收敛精度低，容易陷入局部最优，而HHO算法具有较强的局部开发能力，但存在全局探索能力弱，收敛速度慢的缺陷。针对原始算法存在的局限性，本文将两种算法混合并引入动态反向学习策略，提出一种融合动态反向学习的阿奎拉鹰与哈里斯鹰混合优化算法。首先，在初始化阶段引入动态反向学习策略提升混合算法初始化性能与收敛速度。此外，混合算法分别保留了AO的探索机制与HHO的开发机制，提高算法的寻优能力。仿真实验采用23个基准测试函数和2个工程设计问题测试混合算法优化性能，并对比了几种经典反向学习策略，结果表明引入动态反向学习的混合算法收敛性能更佳，能够有效求解工程设计问题。

Abstract:

In recent years, optimization algorithms such as the Aquila optimizer (AO) and Harris hawks optimization (HHO) have been proposed. Although AO has strong global optimization capabilities, its convergence accuracy is low, and it is susceptible to local optimization. While the HHO algorithm has a strong local development capability, it suffers from flaws such as limited global exploration capabilities and slow convergence speed. Given the limitations of the original algorithms, this paper combines the two algorithms and proposes a hybrid Aquila and Harris Hawks algorithm with dynamic opposition-based learning. It introduces a dynamic opposition-based learning strategy and proposes a hybrid Aquila and Harris Hawks algorithm with dynamic opposition-based learning. First, a dynamic opposition-based learning strategy is introduced in the initialization phase to improve the initialization performance and convergence speed of the algorithm. Second, the hybrid algorithm retains the exploration mechanism of AO and the exploitation mechanism of HHO, which improves the algorithm’s optimization ability. The simulation experiment compares several classical opposition-based learning strategies using 23 benchmark functions and two engineering design problems to test the optimization performance of the hybrid algorithm. The results show that the hybrid algorithm with dynamic opposition-based learning has better convergence performance and can effectively solve engineering design problems.