Abstract: The Grasshopper Optimization Algorithm (GOA), which is based on the grasshopper swarming characteristics, has been demonstrated to effectively solve global constrained optimization problems. However, the original GOA faces challenges, such as an imbalance between exploration and exploitation owing to its linear convergence parameter, slow convergence rate, and tendency to become trapped in local optima. To address these issues, this study proposes a Modified Grasshopper Optimization Algorithm (MGOA). The MGOA incorporates two key modifications: enhanced social interaction and a revised convergence parameter designed to balance exploitation and exploration. The algorithm was evaluated using the CEC2021 benchmark functions, and it outperformed the original GOA in most cases. Statistical analysis using the Wilcoxon signed-rank and Friedman ranking tests further confirmed the significant improvements achieved by the MGOA. These results demonstrate the potential of the MGOA as a more effective optimization approach than the original GOA.

Keywords: Grasshopper optimization algorithm, GOA, meta-heuristics, optimization, swarm intelligence, constrained optimization, CEC2021 benchmark functions, exploration and exploitation, Lévy flight

Pages: 182-198