PRS-P2P: A Prosumer Recommender System for Secure P2P Energy Trading using Q-Learning Towards 6G

In this paper, we present a secure Peer-to-Peer (P2P) trading system for residential houses that aims to maximize energy sharing and improve the operational effectiveness of utility vendors (UVs) in the Smart Grid (SG) environment. Here, we propose PRS-P2P scheme, i.e., a Prosumer Recommender System (PRS) for P2P Energy Trading (ET) using reinforcement learning and blockchain towards 6G. A reinforcement learning-based (Q-learning) algorithm is proposed to improve the decision-making process of the buyer (i.e., consumer) for the seller (i.e., prosumers, who can generate and consume energy) selection from multiple seller participants in a model-free way. During P2P-ET, data flow among different devices, which raises concern for several issues such as security, latency, and others. So, the proposed PRS-P2P scheme uses ethereum blockchain and 6G network, where the buyer sends a request for P2P-ET to the selected seller and trade executed securely using Smart Contract (SC) based on the ethereum blockchain. Then, it employs InterPlanetary File System (IPFS) for energy data management at low cost and performs real-time settlement of trade with low latency. To justify the efficacy of the proposed PRS-P2P scheme, experimental results are compared to the existing approaches concerning different metrics like timesteps for rewards, penalties, and low communication latency.