Optimizing Thermal Comfort and Energy Efficiency in Smart Buildings using GANs and Reinforcement Learning Models

Nearly half of a building's energy comes from the HVAC system. Building energy-efficient thermal management technologies are essential to reduce HVAC energy consumption and ensure occupant comfort. The many environmental variables that affect a building's energy usage are hard to model and change, making such solutions challenging to create. This initiative uses smart technologies to design energy-efficient buildings. Model accuracy is improved by variational dequantization for discrete features and mode-specific normalisation for continuous features. The model was trained for greatest energy savings using RL-GAN. The model had 91.48% accuracy, proving that the proposed method reduces energy use while retaining comfort. This study offers practical advice for creating energy-efficient, occupant-centred smart buildings. The findings show that smart solutions can dramatically reduce energy usage, enabling green HVAC management.