A power scheduling game for reducing the peak demand of residential users

Smart Grids have recently gained increasing attention as a means to efficiently manage the houses energy consumption in order to reduce their peak absorption, thus improving the performance of power generation and distribution systems. In this paper, we propose a fully distributed Demand Management System especially tailored to reduce the peak demand of a group of residential users. We model such system using a game theoretical approach; in particular, we propose a dynamic pricing strategy, where energy tariffs are function of the overall power demand of customers. In such scenario, multiple selfish users select the cheapest time slots (minimizing their daily bill) while satisfying their energy requests. We theoretically show that our game is potential, and propose a simple yet effective best response strategy that converges to a Pure Nash Equilibrium, thus proving the robustness of the power scheduling plan obtained without any central coordination of the operator. Numerical results, obtained using real energy consumption traces, show that the system-wide peak absorption achieved in a completely distributed fashion can be reduced up to 20%, thus decreasing the CAPEX necessary to meet the growing energy demand.