Нүүр зураг

An advanced RIME optimizer with random reselection and Powell mechanism for engineering design

-д хадгалсан:
-д хэвлэсэн:Journal of Computational Design and Engineering vol. 11, no. 6 (Dec 2024), p. 139
Үндсэн зохиолч: Xu, Shiqi
Бусад зохиолчид: Jiang, Wei, Chen, Yi, Ali Asghar Heidari, Liu, Lei, Chen, Huiling, Liang, Guoxi
Хэвлэсэн:
Oxford University Press
Нөхцлүүд:
Qualitative analysis
Algorithms
Convergence
Stability
Performance indices
Sensitivity analysis
Design engineering
Parameter sensitivity
Rank tests
Ablation
Optimization
Онлайн хандалт:Citation/Abstract
Full Text - PDF
Шошгууд: Шошго нэмэх
Шошго байхгүй, Энэхүү баримтыг шошголох эхний хүн болох!
Хураангуй:RIME is a recently introduced optimization algorithm that draws inspiration from natural phenomena. However, RIME has certain limitations. For example, it is prone to falling into Local Optima, thus failing to find the Global Optima, and has the problem of slow convergence. To solve these problems, this paper introduces an improved RIME algorithm (PCRIME), which combines the random reselection strategy and the Powell mechanism. The random reselection strategy enhances population diversity and helps to escape Local Optima, while the Powell mechanism helps to improve the convergence accuracy and thus find the optimal solution. To verify the superior performance of PCRIME, we conducted a series of experiments at CEC 2017 and CEC 2022, including qualitative analysis, ablation studies, parameter sensitivity analysis, and comparison with various advanced algorithms. We used the Wilcoxon signed-rank test and the Friedman test to confirm the performance advantage of PCRIME over its peers. The experimental data show that PCRIME has superior optimization ability and robustness. Finally, this paper applies PCRIME to five real engineering problems and proposes feasible solutions and comprehensive performance index definitions for these five problems to prove the stability of the proposed algorithm. The results show that the PCRIME algorithm can not only effectively solve practical problems, but also has excellent stability, making it an excellent algorithm.
ISSN:2288-5048
DOI:10.1093/jcde/qwae090
Эх сурвалж:Engineering Database