EWO: An Enhanced Walrus Optimizer with Intelligent Reverse Migration and Golden-Sine Strategies for Solving Global and Complex Power and Energy Optimization Problems

This paper proposes an Enhanced Walrus Optimizer (EWO), a novel variant of the recently introduced Walrus Optimizer (WO) for solving global and complex power and energy optimization problems. The original WO, inspired by the cooperative and foraging behaviors of walruses, achieves a reasonable exploration–exploitation balance. However, it is prone to premature convergence and a loss of population diversity in later search stages, limiting its effectiveness on complex problems. To overcome these inherent limitations, EWO integrates two powerful adaptive strategies: novel Intelligent Reverse Migration (IRM) and the Golden-Sine Strategy (GSS). IRM is employed to overcome the inherent limitation of random initialization in the original WO initialization to enhance population diversity and ensure broader exploration of the search space. Concurrently, GSS is embedded within the search dynamics to adaptively balance global exploration and local exploitation, thereby improving convergence accuracy and preventing entrapment in local optima. The performance of EWO is rigorously validated using the CEC-2019 benchmark suite, comprising ten complex functions and four challenging real-world engineering problems: the Traveling Salesman Problem (TSP), Photovoltaic Maximum Power Point Tracking (MPPT) under partial shading, Economic Load Dispatch (ELD), and Microgrid Energy Management. Comprehensive comparative analyses are conducted against the original WO, Particle Swarm Optimization (PSO), Dwarf Mongoose Optimization (DMO), White Shark Optimizer (WSO), and other state-of-the-art algorithms from the literature. Under consistent experimental settings, the results demonstrate that EWO consistently achieves better optimization performance. It yields lower mean fitness values, smaller standard deviations, and faster convergence speeds across all test cases. Statistical validation through Friedman ranking tests and Wilcoxon signed-rank tests confirms the significant and consistent superiority of EWO. The combination of the IRM strategy guarantees a wide global search using the diverse initial population, and GSS provides a smooth changeover from global search to local exploitation using the adaptive step size mechanism. These results show that EWO is a robust, stable, and general-purpose optimization algorithm that can be used to solve complex benchmark functions and complicated power and energy engineering problems.