Exploring harmony search for power system optimization: applications, formulations, and open problems

Modern power systems must optimize large-scale, nonlinear, multi-objective problems created by renewable integration, rapidly growing distributed resources, and strict reliability and efficiency targets. Conventional techniques often falter under these conditions, whereas Harmony Search (HS) has shown strong potential. Unlike earlier HS surveys, this review provides a structured synthesis of power system applications of HS, with attention to how objective functions and constraints are formulated in HS models. It covers both system-level operations (e.g., economic dispatch, optimal power flow, unit commitment, renewable planning) and device-level control (e.g., load frequency regulation and power system stabilization). Comparative results demonstrate HS’s versatility in meeting cost, emission, and reliability goals, and reveal scenarios where careful formulation improves convergence and solution quality relative to other metaheuristics. Practical guidance is distilled on parameter self-adaptation, hybridization with artificial intelligence models, and constraint-handling schemes that mitigate sensitivity and premature convergence. Remaining challenges include inconsistent modeling practices and limited scalability for real-time or very large networks. Recommended remedies encompass standardized HS formulations and automated parameter tuning to improve reproducibility and performance. The review concludes with a future vision of uncertainty-aware, explainable HS frameworks integrated with digital-twin environments, charting a clear agenda for next-generation power-system optimization.