Opposition-Based Dynamic Grey Wolf Optimizer untuk Eksplorasi dan Eksploitasi

Abstract

Grey Wolf Optimizer (GWO) is a prominent swarm intelligence algorithm, but it exhibits significant drawbacks, including a strong search bias towards the origin and premature convergence on complex multimodal landscapes. To address these limitations, this paper proposes a novel hybrid algorithm, the Opposition-Based Dynamic Grey Wolf Optimizer (OB-DGWO). The proposed method integrates a dynamic prey estimation strategy to mitigate search bias with an Opposition-Based Learning (OBL) mechanism to enhance population diversity and global exploration capabilities. The performance of OB-DGWO was rigorously evaluated against the conventional GWO, GWO with OBL, and the dynamic GWO (DGWO) using standard unimodal and multimodal benchmark functions. Experimental results demonstrate that the proposed OB-DGWO exhibits superior robustness. It successfully overcomes the failure of DGWO on problems with optima at the origin, while demonstrating improved accuracy and consistency on complex multimodal functions where the standard GWO fails. The findings indicate that OB-DGWO provides a more balanced and reliable approach for solving diverse optimization problems