A study on pricing strategies for residential load management using fuzzy reinforcement learning

The price based energy consumption control is one of the key approaches for demand side management in smart grids. In this paper, the problem of load control is studied by proposing different pricing strategies from Distribution Company to the users including monthly, daily and hourly price tariffs. The users should decide on their consumption quantity and then time based on pricing strategy of the Distribution Company. The effect of price dynamics in demand side management is also studied. Due to the uncertainties and limited information available to the users, a learning approach is used for decision making of the users in an incomplete information environment. For this purpose, fuzzy Q-Learning method is adapted for learning the optimal policy of the cosumers. In this way, each user learns the best way of using electrical devices to maximize its utility and minimize price. In addition, different pricing mechanisms are compared via simulation results.