|
|
 |
| |
|
|
|
HJSRSO: A Hybrid Algorithm Jellyfish Search and Rat Swarm Optimizer for Engineering Design Problems |
|
|
|
PP: 1813-1838 |
|
|
doi:10.18576/amis/200624
|
|
|
|
Author(s) |
|
|
|
A. M. Rashad,
Abeer A. Ahmed,
Amr Elsonbaty,
M. A. El-Shorbagy,
Moutaz Ramadan,
Hamed A. EL-Mky,
|
|
|
|
Abstract |
|
|
| The study employs a novel technique called HJSRSO, which combines the jellyfish search optimizer (JS) with the rat swarm optimizer (RSO). The Jellyfish Search Optimization (JS) architecture serves as the foundation for the suggested hybrid HJSRSO method. However, the passive motion mechanism (Type A) used in JS is replaced with the Rat Swarm Optimizer (RSO) to successfully enhance the algorithm’s exploitation capabilities. Additionally, a sine-based technique is incorporated into the Jellyfish Search algorithm’s exploration phase to promote exploring behavior. This technique allows agents to follow oscillatory trajectories that boost population diversity and make it easier for the algorithm to escape local optima. Seven constrained engineering design problems (EDPs) and fifty common benchmark functions were used to thoroughly analyze the performance of the suggested HJSRSO algorithm. Rat Swarm Optimizer (RSO), the original Jellyfish Search (JS), and several contemporary cutting-edge metaheuristic algorithms were thoroughly compared. Each algorithm was independently run 30 times under the same experimental conditions for statistical reliability and repeatability. The significance of the suggested modifications to the original JS is demonstrated by the finding that, out of the 50-benchmark functions, HJSRSO obtained the best results in 28 functions with a percentage of 65%, whereas the JS obtained the best solutions in only 18 functions with a percentage of 36%. Key performance metrics, such as the best, worst, mean, median, and standard deviation of the objective function values, were calculated. Non-parametric statistical tests, such as the Friedman rank-based test and the Wilcoxon signed-rank test, were used to quantitatively assess the statistical significance of the observed performance differences. These analyses’ findings verify that HJSRSO regularly achieves better optimal solutions than the original and rival algorithms, demonstrating statistically significant superiority in terms of solution quality, convergence stability, and robustness. |
|
|
|
|
|
|
|
