Explainable Reinforcement Learning for Optimizing Electricity Costs in Building Energy Management

In recent years, the world has shifted its focus towards sustainable energy sources. Efficient usage of renewable sources in buildings requires optimal usage of the fluctuating generation and battery systems. Reinforcement learning (RL) could play a key role in this optimization but mostly remains a black-box in previous applications. In this paper, we propose an Explainable RL (XRL) agent to solve the dynamic problem of cost optimization in energy systems and make it transparent by applying Explainable Artificial Intelligence (XAI) techniques. We demonstrate how the agent learns efficient policies, such as purchasing power at lower prices and selling it at high prices.