Osprey Optimization Algorithm and Position-Aware Graph Neural Network for Enhanced Vehicle-to-Grid Control System

Optimized Vehicle-to-Grid (V2G) Control System enhances energy management by leveraging Electric Vehicle (EV) as flexible storage units, enabling controlled energy exchange with the grid to improve power distribution and overall system efficiency. The system encounters difficulties in reducing total EV charging and discharging plan costs while maintaining high operational efficiency and power grid integrity throughout varying electrical circumstances. To address these challenges, this paper proposes a hybrid approach for optimized V2G control system. The proposed method integrates the Osprey Optimization Algorithm (OOA) and Position-Aware Graph Neural Network (PAGNN); therefore it is called the OOA-PAGNN technique. The main goal is to reduce total costs, enhance grid stability, and generate revenue for EV owners through an optimized V2G framework. OOA is utilized to optimize the charging and discharging schedules of EVs, ensuring effective energy flow management and enhancing grid stability. While, PAGNN is utilized to predict energy demand and generation by analyzing the spatial distribution of EVs, enabling optimized energy allocation. By then, the proposed method is implemented on the MATLAB platform and evaluated with various existing approaches such as Artificial Neural Network (ANN), Mixed-Integer Quadratic Programming (MIQP), Particle Swarm Optimization-Trained Artificial Neural Network (PSO-TANN), Recurrent Proximal Policy Optimization-Long Short-Term Memory (RPPO-LSTM) and Competitive Swarm Optimization-Imperialist Competitive Algorithm (CSO-ICA). The OOA-PAGNN model successfully optimizes V2G control while lowering expenses by -1.458$ while maintaining 98.6% system efficiency.