Enhancing Residential Energy Management with Integration of Photovoltaic Systems using Hybrid Approach

The combination of residential photovoltaic (PV) and battery systems into smart home energy management systems (HEMS) continues to pose several challenges, particularly due to the uncertainty in renewable generation and fluctuations in household energy demand. The capacity of many HEMS models now in use to provide precise forecasting and effective scheduling in a combined framework is limited. To address these limitations, this study gives a novel hybrid approach that combines the Nutcracker Optimization Algorithm (NOA) with Evolutionary Generative Adversarial Networks (EGANs), referred to as the NOA-EGAN framework. In the proposed model, EGAN is used for day-ahead electricity load forecasting. At the same time, NOA is applied to optimize household energy scheduling based on the predicted demand and available solar generation. This integrated method aims to improve forecasting accuracy, enhance cost-effectiveness and enable more adaptive control in PV-powered home energy systems. The proposed approach is implemented and compared in MATLAB with widely used methods like Genetic Algorithm (GA), Ant Lion Optimization (ALO) and Bee Colony Optimization (BCO). Simulation results indicate that NOA-EGAN achieves an energy consumption of 25.57[Formula: see text]kWh, a peak-to-average load ratio of 0.10, and a net-load deviation of 0.89[Formula: see text]kW, all of which are notably lower than those achieved by the comparative methods. The proposed method also demonstrates strong computational performance, attaining the lowest RMSE of 0.321 and the shortest computation time of 0.51 s. Overall, the results suggest that the proposed NOA-EGAN technique is a promising solution for improving cost efficiency and operational performance in smart residential energy management.