A Reliable Energy Trading Strategy in Intelligent Microgrids Using Deep Reinforcement Learning

A microgrid is a system that incorporates various decentralized power sources and manages the distribution of electricity. In microgrid, prosumers play a significant role in trading electricity, and it is essential to establish an effective power trading mechanism to incentivize their active participation. This study proposes a power trading mechanism for prosumers in microgrids, which incorporates blockchain technology to protect their rights and interests. The mechanism employs reinforcement learning to optimize trading decisions and develops a reputation mechanism to evaluate prosumers' trustworthiness based on their past transactions. The proposed strategy encourages prosumers to conduct more transactions with honest peers by introducing reputation value rewards into the benefit function. The simulation results indicate that the proposed strategy outperforms traditional approaches and reputation value significantly impacts prosumers' utility. Overall, the proposed strategy aims to promote honest transactions and enhance prosumers' participation in microgrid power transactions.