Deep Reinforcement Learning and Blockchain for Peer-to-Peer Energy Trading among Microgrids

Abstract

Peer-to-peer energy trading among microgrids has many advantages, e.g., increasing the utilization of renewable energies, reducing the dependence on the main grid and reducing energy cost. In this paper, we investigate a peer-to-peer energy trading problem among microgrids under uncertainties. To be specific, each microgrid intends to maximize its own utility in a local peer-to-peer energy market. Due to the existence of uncertainties in renewable energy output and power demand of all microgrids, and temporally-coupled constraints related to energy storage devices, it is very challenging to develop an optimal energy trading policy for each microgrid. To achieve the above aim, we propose a multiagent deep deterministic policy gradient (MADDPG)-based energy trading algorithm, which can help to find the optimal policy for each microgrid without requiring the generation and load information of other microgrids. Moreover, blockchain is adopted to guarantee the integrity of energy transaction data. Simulation results show the effectiveness of the proposed algorithm in the aspect of reducing energy cost and ensuring the security of transaction data.