Abstract: Real-time electricity price strategy is the effective means to save electricity and improve user electricity utility value. A real-time electricity price optimization model based on Markov Decision Process was raised. Using finite horizon method, the model structure the mathematical model which makes the expected utility maximum of supply side and demand side, and optimize the existing electricity utility function according to decreasing risk theory which using logarithmic form can describe the power utility of user more accurate. Particle Swarm Optimization was used to solve this model and make the results compare with the situation of fixed power price, the results show that this model is better than fixed power price in power saving and utility improving. Beside, the fluctuation of real-time price is between highest price and lowest price, and the fluctuation is not strong.

Key words: smart grid, real-time electricity price, markov decision process, decreasing risk