Back to publications

CntrlDA: A building energy management control system with real-time adjustments

Dmitrewski, Alex and Molina-Solana, Miguel and Arcucci, Rossella
Building and Environment (under review) (SUBMITTED)


Rule-Based Control (RBC) and Model Predictive Control (MPC) have been traditionally used to control building heating, ventilation and air conditioning (HVAC) systems. They, however, present shortcomings when faced with efficiently controlling these systems at a larger level. Reinforcement Learning (RL) has recently emerged as a viable alternative, showing promising results compared to previous methods, but still having some difficulties with untrained situations or sudden changes. CntrlDA is our proposal on improving the RL formulation by coupling it with Data Assimilation (DA), a technique commonly used in numerical weather prediction. Our battery of experiments, in a building simulation environment, shows that training a RL control agent with DA and external data, leads to better performance than training the agent using only the simulation data. It is also shown that by including this DA stage, the agent better learns how to deal with unexpected events, which are common in real-life systems, and particularly in building energy control scenarios.




  author = {Dmitrewski, Alex and Molina-Solana, Miguel and Arcucci, Rossella},
  title = {CntrlDA: A building energy management control system with real-time adjustments},
  journal = {Building and Environment},
  year = {SUBMITTED},
  volume = {(under review)},
  doi = {},
  comment = {},
  timestamp = {33}