Game theoretic model predictive control for distributed energy demand-side management

Abstract

Distributed energy generation and storage are widely investigated demand-side management (DSM) technologies that are scalable and integrable with contemporary smart grid systems. However, prior research has mainly focused on day-ahead optimization for these distributed energy resources while neglecting forecasting errors and their often detrimental consequences. We propose a novel game theoretic model predictive control (MPC) approach for DSM that can adapt to real-time data. The MPC-based algorithm produces subgame perfect equilibrium strategies for distributed generation and storage with perfect forecasting information, and is shown to be more effective than a day-ahead scheme when mean forecasting errors greater than 10% are present. This robust and continuous MPC approach reduces effective forecasting errors, and in doing so, achieves greater electricity cost savings and peak to average demand ratio reduction than the day-ahead optimization scheme.