TY - JOUR

T1 - DGS-SCSO: Enhancing Sand Cat Swarm Optimization with Dynamic Pinhole Imaging and Golden Sine Algorithm for improved numerical optimization performance

AU - Adegboye, Oluwatayomi

AU - Feda, Afi

AU - Ojekemi, Oluwaseun Racheal

AU - Agyekum, Ephraim

AU - Khan, Baseem

AU - Kamel, Salah

PY - 2024/1

Y1 - 2024/1

N2 - This paper introduces DGS-SCSO, a novel optimizer derived from Sand Cat Swarm Optimization (SCSO), aiming to overcome inherent limitations in the original SCSO algorithm. The proposed optimizer integrates Dynamic Pinhole Imaging and Golden Sine Algorithm to mitigate issues like local optima entrapment, premature convergence, and delayed convergence. By leveraging the Dynamic Pinhole Imaging technique, DGS-SCSO enhances the optimizer's global exploration capability, while the Golden Sine Algorithm strategy improves exploitation, facilitating convergence towards optimal solutions. The algorithm's performance is systematically assessed across 20 standard benchmark functions, CEC2019 test functions, and two practical engineering problems. The outcome proves DGS-SCSO's superiority over the original SCSO algorithm, achieving an overall efficiency of 59.66% in 30 dimensions and 76.92% in 50 and 100 dimensions for optimization functions. It also demonstrated competitive results on engineering problems. Statistical analysis, including the Wilcoxon Rank Sum Test and Friedman Test, validate DGS-SCSO efficiency and significant improvement to the compared algorithms.

AB - This paper introduces DGS-SCSO, a novel optimizer derived from Sand Cat Swarm Optimization (SCSO), aiming to overcome inherent limitations in the original SCSO algorithm. The proposed optimizer integrates Dynamic Pinhole Imaging and Golden Sine Algorithm to mitigate issues like local optima entrapment, premature convergence, and delayed convergence. By leveraging the Dynamic Pinhole Imaging technique, DGS-SCSO enhances the optimizer's global exploration capability, while the Golden Sine Algorithm strategy improves exploitation, facilitating convergence towards optimal solutions. The algorithm's performance is systematically assessed across 20 standard benchmark functions, CEC2019 test functions, and two practical engineering problems. The outcome proves DGS-SCSO's superiority over the original SCSO algorithm, achieving an overall efficiency of 59.66% in 30 dimensions and 76.92% in 50 and 100 dimensions for optimization functions. It also demonstrated competitive results on engineering problems. Statistical analysis, including the Wilcoxon Rank Sum Test and Friedman Test, validate DGS-SCSO efficiency and significant improvement to the compared algorithms.

UR - http://www.scopus.com/inward/record.url?partnerID=8YFLogxK&scp=85182441659

U2 - 10.1038/s41598-023-50910-x

DO - 10.1038/s41598-023-50910-x

M3 - Article

VL - 14

JO - Scientific Reports

JF - Scientific Reports

SN - 2045-2322

IS - 1

M1 - 1491

ER -