Supplementary material to the paper. Includes datasets, charts and additional code.

The experiments were conducted using Sinergym, a framework for building energy simulation.

Heating, ventilation, and air conditioning (HVAC) systems are a major driver of energy consumption in commercial and residential buildings. Recent studies have shown that Deep Reinforcement Learning (DRL) algorithms can outperform traditional reactive controllers. However, DRL-based solutions are generally designed for ad hoc setups and lack standardization for comparison. To fill this gap, this paper provides a critical and reproducible evaluation, in terms of comfort and energy consumption, of several state-of-the-art DRL algorithms for HVAC control. The study examines the controllers’ robustness, adaptability, and trade-off between optimization goals by using the Sinergym framework. The results obtained confirm the potential of DRL algorithms, such as SAC and TD3, in complex scenarios and reveal several challenges related to generalization and incremental learning.

Reinforcement Learning, HVAC, Building Energy Optimization, Sinergym

• Antonio Manjavacas
• Alejandro Campoy-Nieves
• Javier Jiménez-Raboso
• Miguel Molina-Solana
• Juan Gómez-Romero

Department of Computer Science and Artificial Intelligence (DECSAI), Universidad de Granada, Spain.

Sustainable Artificial Intelligence Lab (SAIL).

@article{manjavacas2024experimental,
  title={An experimental evaluation of deep reinforcement learning algorithms for HVAC control},
  author={Manjavacas, Antonio and Campoy-Nieves, Alejandro and Jim{\'e}nez-Raboso, Javier and Molina-Solana, Miguel and G{\'o}mez-Romero, Juan},
  journal={Artificial Intelligence Review},
  volume={57},
  number={7},
  pages={173},
  year={2024},
  publisher={Springer}
}

Read LICENSE.