Distributed Optimal Dispatch Method for Smart Community Demand Response Based on Machine Learning

With the development of power system, accurate demand response business is developing towards diversification, which requires higher reliability and accuracy of demand response. In recent years, machine learning has been a hot topic in the academic field. If the ability of machine learning to learn from the environment and make optimal decisions is applied to the demand response, the demand response business can have a deep understanding of users behavior and make accurate predictions. Combining with the concept of smart community, this paper proposes a distributed optimal dispatch method of demand response based on LSTM and Q-learning algorithm. In this paper, the algorithm is applied to the demand response business planning of an smart community in a city with tropical monsoon climate. The simulation results show that the algorithm is consistent with the facts and has high accuracy.