Optimizing energy regulation of electric vehicles in incentive-based prosumer microgrids with uncertainty modeling

Community microgrids have emerged as a sustainable and decentralized energy solution, facilitating local generation, storage, and distribution while enhancing grid resilience and reducing dependency on conventional power sources. However, their effective operation is hindered by the dynamic nature of energy generation and consumption. This work introduces a novel methodology for optimizing the charge/discharge scheduling of Electric Vehicles (EVs) within a community microgrid environment. The proposed approach, termed MSNN-GBO, combines a Mix Style Neural Network (MSNN) for forecasting EV availability with the Gooseneck Barnacle Optimization (GBO) algorithm for efficient resource allocation under uncertainties such as fluctuating EV availability and dynamic electricity pricing. The framework integrates the operational characteristics of EVs, photovoltaic (PV) systems, and energy storage systems (ESS), while accounting for battery degradation. The aim is to reduce operational costs and enhance economic performance for prosumers. Implemented in MATLAB, the proposed technique is compared with the Alternating Direction Method of Multipliers Algorithm (ADMMA), Binary Dragonfly Algorithm (BDA), and Proximal Policy Optimization (PPO). Simulation results reveal cost reduction of up to 19.4%, decreasing from 11.25$ to 9.42$, and achieving a system efficiency of 98%. The findings underscore the proposed method's potential to improve reliability, cost-efficiency, and sustainability in community microgrid operations.