Efficient Algorithms and Policies for Demand Response Scheduling

Abstract

The authors consider efficient mechanisms to optimize the power consumption within a home, industrial facility, college campus, or other facility or set of facilities. The system is controlled centrally by an energy management controller (EMC), which determines the timing of the operation of some of the devices within the facilities. The authors introduce an approximate dynamic programming (ADP) algorithm for this problem and show that the ADP outperforms a recent dynamic programming (DP) algorithm. However, even the ADP fails to solve sufficiently quickly when applied to larger instances. Therefore, the authors also propose several heuristic scheduling policies that provide accurate solutions in a fraction of the time required by the ADP. The authors discuss the computational performance of the ADP algorithm and scheduling policies, and insights gained from the models.