Optimal control of smart home energy management based on stochastic dynamic programming

The optimal energy management problem of smart home is concerned. The objective is to maximize reward of smart home with distributed generation, through choosing a suitable energy management control strategy by stochastic dynamic programming. Firstly, we modeled the smart home energy management system as a semi-Markov decision process according to the stochastic characteristics of solar photovoltaic, load demand and electricity price. We apply policy iteration algorithm to solve the optimal control of energy management. Then, SARSA algorithm based on simulated annealing is applied to optimize the energy management strategy for the system. Finally, a simulation example is used to illustrate the effectiveness of the algorithm, and the results show that the algorithm can effectively increased the user rewards.