Reinforcement learning for the control of building energy systems

One funded Master’s or Ph.D. position is available starting now in the Department of System Engineering at École de technologie supérieure (ÉTS) in Montréal. The student will work under the supervision of Prof. Fausto Errico. Fausto Errico is also a member of CIRRELT and GERAD, two internationally recognized research centers in Operations Research and Machine Learning located in Montréal. Opportunities for co-supervision with other researchers from these centers or from ÉTS are possible and welcome.

The research project focuses on the use of reinforcement learning as a central methodology for the optimization and control of building energy systems, with HVAC systems as the main actuation layer. As buildings account for a significant share of global energy consumption and emissions, their interaction with power grids increasingly dominated by renewable energy sources raises new control challenges. In this context, buildings are expected to move from passive energy consumers to active systems capable of adapting their energy demand under uncertainty, while preserving indoor comfort and occupant well-being.

The research will explore hybrid control architectures combining classical controllers (e.g., PID) with data-driven and learning-based decision layers. Topics may include adaptive tuning of control parameters, dynamic adjustment of temperature targets based on occupancy and context, and the integration of uncertainty and forecasts related to external factors such as weather and usage patterns.

The project combines methodological developments with realistic building control scenarios. It is carried out in collaboration with a major industrial actor in building automation and energy management, providing strong connections to practical control problems and operational constraints.

Candidates should have a background in Operations Research, Applied Mathematics, Computer Science, Control, Machine Learning, or a related field. A strong interest in optimization, control, reinforcement learning, and learning-based methods is expected. Experience with Python and/or C/C++ programming is required, and familiarity with reinforcement learning, optimization, control, or machine learning techniques is an asset.

Interested candidates should send their application to Fausto Errico (fausto.errico@etsmtl.ca). Applications should be submitted as a single PDF file and include:

1. A short cover letter describing the candidate’s background, motivation, and preferred start date
2. A curriculum vitae
3. Academic transcripts
4. A copy of one or two representative papers or technical reports, if applicable
5. (For Ph.D. applicants) A copy of the Master’s thesis, if available
6. Contact information for one or two academic referees

Applications may be written in English or French. Fluency in English is required; knowledge of French is an asset but not mandatory.

Review of applications will begin immediately and continue until the positions are filled.