Using smart devices for system-level management and control in the smart grid: A reinforcement learning framework

This paper presents a stochastic modeling framework to employ adaptive control strategies in order to provide short term ancillary services to the power grid by using a population of heterogenous thermostatically controlled loads. The problem is cast anew as a classical Markov Decision Process (MDP) to leverage existing tools in the field of reinforcement learning. Initial considerations and possible reductions in the action and state spaces are described. A Q-learning approach is implemented in simulation to demonstrate how the performance of the new MDP representation is comparable to that of a Linear Time-Invariant (LTI) one on a reference tracking scenario.