Journal of Frontiers of Computer Science and Technology ›› 2023, Vol. 17 ›› Issue (9): 2118-2136.DOI: 10.3778/j.issn.1673-9418.2203134

• Theory·Algorithm • Previous Articles     Next Articles

HHO Algorithm Based on Bidirectional Experience Guidance and Extreme Individual Regulation

CHAI Yan, REN Sheng   

  1. College of Science, Liaoning Technical University, Fuxin, Liaoning 123000, China
  • Online:2023-09-01 Published:2023-09-01

双向经验引导与极端个体调控的HHO算法

柴岩,任生   

  1. 辽宁工程技术大学 理学院,辽宁 阜新 123000

Abstract: In order to improve the optimization accuracy and iteration speed of Harris hawks optimization (HHO) algorithm, a bidirectional experience guided and extreme individual regulated HHO algorithm (BEHHO) is proposed. Firstly, Circle chaotic mapping is used to homogenize the initial population to effectively avoid individual aggregation and improve the coverage of the Harris hawks population to the solution space region, laying foundation for algorithm optimization. Secondly, the bidirectional experience guidance strategy is introduced to strengthen the rounding mechanism of the algorithm, the evolution experience of the global optimal individual and the historical optimal individual is used to guide the individual to search for the optimal direction, and the differential disturbance term of the adaptive random individual is used to strengthen the ability of the population to explore the neighborhood and improve the convergence accuracy of the algorithm. Furthermore, considering important impact of algorithm extreme individual on the global update process, the optimal individual t - distribution variation is used to avoid algorithm into local extremum, and the reverse solution of the worst individual is generated by reverse dynamic learning to improve the convergence speed indirectly. At the same time, greedy principle is adopted to retain the dominant individual to ensure the accuracy of the algorithm’s progeny individual tends to be better. Finally, the global convergence of the algorithm is analyzed based on Markov chain. By comparing the optimization of benchmark test functions, Wilcoxon rank sum test and CEC2014 complex functions, the improved algorithm is proven to have excellent solving performance and robust robustness, and the superiority of BEHHO algorithm in practical problems is verified by welding beam design problems in engineering optimization.

Key words: Harris hawks optimization (HHO) algorithm, Circle chaotic mapping, bidirectional experience guidance, extreme individual regulation, global convergence, engineering optimization

摘要: 为进一步提升哈里斯鹰优化算法(HHO)的寻优精度和迭代速度,提出一种双向经验引导与极端个体调控的HHO算法(BEHHO)。首先采用Circle混沌映射均匀化初始种群,有效规避个体聚集情形并提升哈里斯鹰群体对解空间区域的覆盖性,奠定算法寻优基础;其次引入双向经验引导策略来强化算法的围捕机制,依托全局最优个体和历史最优个体的进化经验引导个体寻优方向,且配合自适应随机个体的差分扰动项来强化种群探索邻域能力,提升算法的收敛精度;再者考虑算法中极端个体对全局更新过程的重要影响,利用t-分布变异最优个体来避免算法陷入局部极值区,并以动态反向学习产生最差个体的反向解来间接提高算法的收敛速度,同时采用贪婪原则保留优势个体的方式确保算法子代精度趋于更优;最后基于马尔科夫链分析算法的全局收敛性。通过对基准测试函数的寻优对比分析、Wilcoxon秩和检验以及CEC2014复杂函数的对比分析,验证了改进算法优异的求解性能和健壮的鲁棒性,并以工程优化中焊接梁设计问题验证了BEHHO算法处理实际问题时的优越性。

关键词: 哈里斯鹰优化算法(HHO), Circle混沌映射, 双向经验引导, 极端个体调控, 全局收敛性, 工程优化