Deep Reinforcement Learning Enabled Incentive Mechanism of Electric Vehicles for Renewable Energy Power Transmission
摘要
Renewable energy power has advantages such as sustainability and cleanliness, which helps to save energy and reduce emissions, improve energy utilization efficiency, and protect the environment. Nevertheless, it exerts significant pressure on power grids when large scale renewable energy power integrates into power grid straightforwardly, such as the safe and stable operation and economic efficiency of the power grid system. To resolve the above power grid challenges, we propose a renewable energy power transmission system which overlays Electric Vehicles (EV), for the sparse deployment of renewable energy power stations, combing the Vehicle-to-Grid technology. We formulate the Social optimization EV User Selection (SEUS) problem and show that the SEUS problem is NP-hard. We design the incentive mechanism based on Deep Reinforcement Learning and greedy approach. By means of rigorous theoretical analysis and empirical simulations, we have established that our proposed incentive mechanism can ensure individual rationality and truthfulness, and significantly surpass the benchmark algorithms.