Reinforcement Learning Based Peer-to-Peer Energy Trade Management Using Community Energy Storage in Local Energy Market

Many studies have proposed a peer-to-peer energy market where the prosumers’ actions, including consumption, charge and discharge schedule of storage systems, transactions in local markets, are controlled by central operator. In this paper, actions not an operator, prosumers freely participate to trade with other prosumers. We designed modeled management algorithm that uses community for who competitively real-time market. propose energy-trade manages trades two phases based on bids offers submitted The first phase is manage fair prices second managing could phase. Community employed reinforcement learning-based trading decide whether buy, sell, or do nothing action buying selling means charging discharging storage, respectively. Numerical results show gains near-maximum profit. Besides, we verified yields more profit than battery wear-out cost.