Hybrid Optimization in Prosumer-to-Grid Energy Management System for Pareto-Optimal Solution

Energy management in smart microgrids is critical to achieving sustainable and efficient energy utilization. This study introduces a hybrid optimization framework combining neural networks (NNs) and multi-objective genetic algorithms (MOGAs) (hybrid NN-MOGA) to address the dual objectives of minimizing total energy cost and maximizing customer satisfaction. The hybrid NN-MOGA approach leverages NNs for predictive modeling of load and renewable energy generation, feeding accurate inputs to the MOGA for enhanced Pareto-optimal solutions. The performance of the proposed method is benchmarked against traditional optimization techniques, including MOGA, multi-objective particle swarm optimization (MOPSO), and the multi-objective firefly algorithm (MOFA). The simulation results demonstrate that hybrid NN-MOGA outperforms the alternative model. The proposed method produces uniformly distributed and highly convergent Pareto frontiers, ensuring robust trade-offs of USD 48.2817 and 81.7898 for total cost and customer satisfaction, respectively. Convexity analysis and the satisfaction of Karush–Kuhn–Tucker (KKT) conditions further validate the optimization model.