Optimal Peer-to-Peer Energy Trading Using Machine Learning: Architecture, Strategies, and Algorithms

Abstract

In this chapter, we discuss the development of machine-learning applications to obtain optimal peer-to-peer (P2P) energy trading. P2P energy trading is developed to facilitate direct electricity transactions among end-node customers. However, the P2P energy trading operation is affected by various challenges, from utilizing big data owing to generation and load data, the behavior of sellers and buyers, and the complexity in obtaining optimal matching decisions. To reduce the effects of these challenges, machine-learning techniques have been applied to P2P energy trading. Machine learning advances the data analysis process and thus yields an optimal pattern that contains important features regarding the overall datasets. Thus, this pattern is considered to enhance the trading mechanism of P2P. A case study of hybrid P2P energy trading via a clustering approach is simulated to demonstrate the advancement afforded by machine-learning applications.