The University of Auckland
Browse

The utilization of machine learning techniques in the building design stage: a qualitative review

Download (854.93 kB)
Version 2 2021-02-02, 04:03
Version 1 2021-01-15, 02:41
conference contribution
posted on 2021-02-02, 04:03 authored by Mustika Sari, Mohammed Ali Berawi

This item is part of: Boarin, P., Haarhoff, E., Manfredini, M., Mohammadzadeh, M., Premier, A., (2021). Rethinking Sustainable Pacific Rim Territories. Proceedings of the 2020 APRU Sustainable Cities and Landscapes Hub PhD Symposium, Future Cities Research Hub, School of Architecture and Planning of the University of Auckland. ISBN: 978-0-473-53616-9


ABSTRACT

Machine learning (ML) is a subdivision of artificial intelligence (AI) technology that has been extensively researched and applied in the building and construction industry for the past few years, particularly in the building design process. It yields significant benefits to the sector through its learning method that not only offers automation for repetitive complex design tasks and evaluation for decision-making functions but also optimization of building performances. This study investigates the existing exploration of the machine learning techniques utilized in the building design process to improve both the performance and result of the building design in the early design stage. From the 27 publications discussed in this paper, the Genetic Algorithm was the most utilized technique in the building design research area, followed by Neural Network. However, the employment of machine learning in this area mostly focused on particular building aspects, creating an opportunity for further studies to develop comprehensive building design solutions through the machine learning approach

History

Publisher

Future Cities Research Hub, School of Architecture and Planning of the University of Auckland