Real-time pricing considering different type of smart home appliances based on Markov decision process

Abstract

Demand side management (DSM) aims to match power demand to supply through cutting the peak and filling the valley. It is one of the most important factors in smart grid. The real-time pricing (RTP) scheme is an ideal method to adjust power balance between supply and demand. Considering the classification of smart home appliance (SHA) and the correlation of power consumption among users, a social welfare maximization model is proposed based on Markov decision process (MDP) in the research. A probabilistic transfer matrix has been introduced to characterize the elastic appliance in smart home. Several state transfer functions have been utilized to reflect the operating process of the semi-elastic appliance also. In this way, the specific characteristics of SHA can be fully embodied. In order to improve computing efficiency and protect the user’s privacy, this project divides the optimization into two subproblems in terms of users and energy supplier (ES). At user side, a modified simulated annealing RTP algorithm is developed to solve the optimization problem. Also, at ES side, the dual sub-gradient method is used to solve the convex optimization problem in the paper. Finally, the simulation results validate the rationality and feasibility of the optimization model by the decentralized RTP algorithm.