Differentially Dynamic Pricing for Local Energy Provider with Renewable Sources

In this paper, we study the dynamic pricing problem for a local energy provider with renewable sources in smart grid. In particular, the local energy provider is connected to the main grid and multiple customers with different flexible loads. Our goal is to maximize social welfare (the weighted sum of the interests of local power provider and multiple end users) by developing differentiated pricing strategies. The challenges of solving optimal pricing decisions here are mainly divided into three aspects: 1) the demand of the end users can respond to the price in real time; 2) the end users unmet elasticity load is shifted to the next time slot, which is a time-coupled constraint; 3) the uncertainty of renewable energy generation, end users demand and external grid price. To address the challenges of dynamic pricing, we propose a dynamic pricing algorithm based on Q-learning framework, which enables local power supplier to understand the behavior of end users and changes in energy costs, and thus determine optimal pricing strategies. The simulation results show that the dynamic pricing algorithm proposed can effectively reduce the cost and dissatisfaction level of end users, while improving the profit of local power supplier and the stability of the power system.