Comparative Assessment of Optimization Strategies with a Hybrid Branch-and-Cut Time Decomposition for Optimal Energy Management Systems

The integration of electric vehicles into microgrids demands advanced energy management to coordinate charging with renewable generation and storage resources. This study presents a cohesive and comprehensive evaluation of four distinct optimization strategies—genetic algorithm (GA), particle swarm optimization (PSO), ant colony optimization (ACO), and mixed-integer linear programming (MILP)—in coordinating EV charging and energy dispatch within a 55 MW grid-connected microgrid that includes photovoltaic, wind, battery energy storage (BESS), and bidirectional EV systems. Beyond numerical outcomes, this work emphasizes the behavioral and methodological characteristics of each optimization approach, assessing their structural advantages and resource utilization dynamics. A novel MILP solution algorithm is introduced, based on a hybrid branch-and-cut technique integrated with time decomposition, enabling the solver to capture long-horizon optimization dynamics with high precision. All four methods are applied over a year-long simulation with hourly resolution. While each strategy maintains operational feasibility and power balance, the MILP approach consistently achieves the highest economic benefit, delivering approximately $2.43 million in annual cost savings, representing roughly a 72.3% improvement over the best-performing heuristic strategy under the same deterministic operating conditions. GA, PSO, and ACO each capture moderate benefits but show limitations in foresight and storage cycling. The findings not only benchmark algorithmic performance but also provide insight into the internal logic and structural behavior of optimization techniques applied to dynamic energy systems, offering guidance for algorithm selection and design in microgrid EMS.