Optimal vehicle-to-grid control for supplementary frequency regulation using deep reinforcement learning

•This research focus on the potential benefits of electric vehicles battery storage and vehicle-to-grid (V2G) technology to provide grid frequency response services.•We propose an optimal V2G control based on deep reinforcement learning which maximises profit to both electric vehicle (EV) owners and the energy aggregator.•The approach uses the deep deterministic policy gradient (DDPG) for scheduling the EV batteries charging/discharging to simultaneously satisfy the driving demand and participate in frequency regulation.•The proposed control scheme has been successfully tested on a two-area interconnected power system undergoing frequency deviations.