Multi-strategy improved artificial rabbit optimization algorithm based on fusion centroid and elite guidance mechanisms
Hefan Huang,Rui Wu,Haisong Huang,Jianan Wei,Zhenggong Han,Long Wen,Yage Yuan
DOI: https://doi.org/10.1016/j.cma.2024.116915
IF: 6.588
2024-03-31
Computer Methods in Applied Mechanics and Engineering
Abstract:The Artificial Rabbit Optimization (ARO) algorithm has been proposed as an effective metaheuristic optimization approach in recent years. However, the ARO algorithm exhibits shortcomings in certain cases, including inefficient search, slow convergence, and vulnerability to local optima. To address these issues, this paper introduces a multi-strategy improved Artificial Rabbit Optimization (IARO) algorithm. Firstly, in the enhanced search strategy, we propose integrating the centroid guidance mechanism and elite guidance mechanism with the greedy strategy to update the position during the exploration phase. Additionally, the Levy flight strategy integrated with self-learning, is employed to update the position during the development phase to improve convergence speed and prevent falling into local optima. Secondly, the algorithm incorporates a per-dimension mirror boundary control strategy to map individuals exceeding the boundary back within the boundary back inside the boundary. This boundary control strategy ensures the algorithm operates within bounds and enhances convergence speed. Finally, within the survival of the fittest strategy, an adaptive factor is introduced to gradually enhance the population's overall adaptability. This factor regulates the balance between exploration and exploitation, allowing the algorithm to fully explore the search space and improve its robustness. To substantiate the effectiveness of the proposed IARO algorithm, a rigorous and systematic verification analysis was undertaken. Comparative experiments for qualitative and quantitative analysis were conducted on three benchmark test sets, namely CEC2017, CEC2020, and CEC2022. The analysis results, including the Wilcoxon rank-sum test, consistently demonstrates that this improved algorithm outperforms ARO and other state-of-the-art optimization algorithms comprehensively. Finally, the feasibility of the IARO algorithm has been verified in seven classical constrained engineering problems.
mechanics,engineering, multidisciplinary,mathematics, interdisciplinary applications