Exploring Deep Generative Models in Building Design

Soheila Kookalani, Erika Parn, Ioannis Brilakis, Ning Pan, Mudan Wang, Hamidreza Alavi
DOI: 10.35490/EC3.2024.235
Abstract: The complex and time-consuming nature of building design necessitates meticulous attention to detail and adherence to principles. Automation in the design process is crucial with a growing demand for innovative and sustainable structures. This review explores the transformative impact of integrating deep generative models into building design, showcasing their ability to generate realistic 3D models, layouts, and structural designs. These models address challenges in architectural design and sustainability assessment, enabling generative design and energy-efficient building design. The review suggests potential solutions and highlights the role of the models in enhancing sustainability, cost-efficiency, and creativity in the built environment.
Keywords: Deep generative models, Design Automation, Generative adversarial networks, Reinforcement Learning, Variational autoencoders
Download paper

Presentation video

Successfully submitted

Your submission has been received. We will review your details and contact you soon.