Back to publications

SINERGYM - A virtual testbed for building energy optimization with Reinforcement Learning

Campoy-Nieves, Alejandro and Manjavacas, Antonio and Jiménez-Raboso, Javier and Molina-Solana, Miguel and Gómez-Romero, Juan
Energy and Buildings (under review) (SUBMITTED)

Abstract:

Simulation has become a crucial tool for Building Energy Optimization (BEO) as it enables the evaluation of different design and control strategies at a low cost. Machine Learning (ML) algorithms can leverage large-scale simulations to learn optimal control from vast amounts of data without supervision, particularly under the Reinforcement Learning (RL) paradigm. Unfortunately, the lack of open and standardized tools has impeded the widespread application of ML and RL to BEO. To address this issue, this paper introduces SINERGYM, an open-source virtual testbed for building simulation, data collection, control, and monitoring. The software is written in Python and backed by the EnergyPlus engine. SINERGYM provides a consistent interface for training and running controllers, predefined benchmarks, experiment visualization and replication support, and comprehensive documentation in a ready-to-use software library. This paper describes the main features of SINERGYM compared to other existing frameworks and its basic use, showing its simplicity and potential for RL-based BEO research and development.

Links:

DOI:
PDF:

Bibtex:

@article{Campoy2023,
  author = {Campoy-Nieves, Alejandro and Manjavacas, Antonio and Jiménez-Raboso, Javier and Molina-Solana, Miguel and Gómez-Romero, Juan},
  title = {SINERGYM - A virtual testbed for building energy optimization with Reinforcement Learning},
  journal = {Energy and Buildings},
  year = {SUBMITTED},
  volume = {(under review)},
  doi = {},
  comment = {},
  timestamp = {38}
}