🚨New paper published by members of our team including Mr Bingnan Zhao, Dr Harshala Gammulle, Dr Tharindu Fernando, Prof Clinton Fookes, and D/Prof Lidia Morawska.
This paper titled, “Optimizing Indoor Air Quality in HVAC Systems Using Deep Reinforcement Learning with Multi-Objective Dynamic Weight Reward Function” is published in Building and Environment.
With a growing emphasis on building occupants’ health and well-being, ensuring better IAQ while simultaneously managing thermal comfort and energy conservation has become a fundamental challenge for intelligent HVAC control. Existing Deep Reinforcement Learning (DRL) approaches predominantly rely on fixed-weight reward functions that cannot adapt to dynamically changing occupancy conditions, and few studies treat IAQ as a primary optimization objective. This study proposes a novel DRL framework whose core innovation is a CO2-driven reward shaping mechanism that couples IAQ sensing with multi-objective control.
Read the paper 👉 https://doi.org/10.1016/j.buildenv.2026.114917
QUT (Queensland University of Technology), Australian Research Council (ARC)
