Real-Time Pricing for Demand Response Based on Stochastic Approximation

In this paper, we propose a new pricing algorithm to minimize the peak-to-average ratio (PAR) in aggregate load demand. The key challenge that we seek to address is the energy provider's uncertainty about the impact of prices on users' load profiles, in particular when users are equipped with automated energy consumption scheduling (ECS) devices. We use an iterative stochastic approximation approach to design two real-time pricing algorithms based on finite-difference and simultaneous perturbation methods, respectively. We also propose the use of a system simulator unit (SSU) that employs approximate dynamic programming to simulate the operation of the ECS devices and users' price-responsiveness. Simulation results show that our proposed real-time pricing algorithms reduce the PAR in aggregate load and help the users to reduce their energy expenses.