A Systematic Review of AI-Powered Generative Design to Improve Efficiency and Thermal Comfort in Early-Stage Residential Design
摘要
The generative design for energy efficiency and thermal comfort in residential buildings is a growing field with increasing research interest. The review identified diverse generative design methods such as EPSAP, machine learning, and rule-based approaches, often integrated with advanced software to optimize design variables such as window characteristics and insulation. Studies have achieved significant improvements in energy efficiency (up to a 13% reduction) and occupant comfort (up to a 38% increase), while highlighting the importance of location-specific design and the impact of urban context on building performance. Additionally, the review found success in combining natural ventilation with conventional air conditioning for substantial energy savings (18% to 40%). The advancements in generative design in optimizing energy efficiency and thermal comfort go beyond mere building performance. They signal a significant shift in contemporary architecture – the growing importance of embedding sustainability principles directly within the design process itself.