Title: Holistic Control for Building Energy Management using Deep Reinforcement Learning
Speakers: Nuno Dionísio, LASIGE – DI/FCUL
When: March 24, 14h00
Where: C6.3.27

Abstract: Given the need for more efficient energy management while maintaining comfort in both residential and commercial buildings, there has been focused effort in designing better control strategies. These often include the control of Heating, Ventilation, and Air Conditioning (HVAC) systems, blind systems, window systems, and lighting systems. However, most existing controllers that are currently deployed in smart buildings mainly follow heuristic-based approaches and are not able of taking into account all the control systems available. This presentation will feature a brief overview of a data-driven holistic approach to energy management in smart buildings that is mainly based on Deep Reinforcement Learning (DRL) techniques.

Short bio: Nuno Dionísio is a PhD candidate in Computer science and also a Machine learning researcher under the EU SATO and Smart2B projects, currently working on his thesis under the supervision of Professor Pedro Ferreira.