Battery Control in a Smart Energy Network using Double Dueling Deep Q-Networks

In this paper, we use the deep reinforcement learning algorithm Double Dueling Deep Q-Networks to manage an energy storage system in a smart energy network. The network models a microgrid with its own demand and renewable energy generation based on a novel dataset collected at Keele University, as well as using dynamic energy pricing from a real wholesale energy market. The algorithm outperforms a model-based benchmark as well as two other Deep Q-Network variants and can be used as the basis of future research.