Maremat & Me'mari-e Iran
Art University of Isfahan
1- , prof.farkish@gmail.com
Abstract: (1267 Views)
Before the emergence of the concept of generative design, AI algorithms did not have a significant presence in the design process. Specially in the early phases when the architect is faced with a vague and foggy image. They were mostly used in 2D and 3D software and accompanied architect in the form of completing his initial ideas. But in recent years, following the expansion of this concept, the presence of them in the design process has been faced some changes. Today the algorithms try to provide all possible answers to the architectural problems to clarify the ambiguous path of design for the architect and help him generate various ideas, especially in the beginning of the process. Many studies have used algorithms in different issues in generative design process, but the majority of them indicate the extent and dispersion in the use of this new method. Based on this premise and in order to achieve a comprehensive view, this research aims to review the state-of-the-art of the algorithms application in AI-based generative design process. The focus is on investigating the studies from the past three decades, that have used this method in the early stages of the design process. First an exhaustive search for literature on the subject has been carried out through the scientific portals ScienceDirect, Google Scholar, Scopus, ResearchGate and IEEE, ACM website. The keywords used were the combination key word in AI and Architecture such as (‘artificial intelligence’, Or ‘AI’ ) Or (‘AI ‘And ‘neural network’ ) Or ( ‘AI’ And ‘machine learning’), Or, (‘AI’ And ‘deep learning’), Or (‘generative design process’) Or (‘automation’) And (‘evolutionary algorithm’) And (‘architectural idea Or concept’) Or (‘architectural design process’) Or (‘architectural form design’ Or ‘ floor plan’ Or ‘ space layout’). After removing articles that were not compatible with the research purpose, a total of 540 articles were subjected to text analysis. Finally, 73 articles were selected based on two criteria: they practically used algorithms in their study, and they utilized them in the initial stages of the process. For the purpose of organizing and analyzing the content of the selected studies more accurately, the topics used in these researches have been divided into six thematic classifications which are based on literature and the final selected papers analysis. There are: idea and concept, form-finding, building shape, roof and facade design, layout design, and plan design. The results indicate that over 75% of these studies independently utilized evolutionary algorithms in the early design phases, with their focus on finding the optimal idea, concept, or spatial layout. While in recent years with the advent of deep learning algorithms, the tendency of researchers to use this type of algorithm has increased. The topics in which they have widely used are designing architectural shape and plan. The combination of these two is also widely used in recent researches. It should be noted that evolutionary algorithms, are still popular among researchers and are widely used independently in generative design process
Article number: 155
Keywords: design process, generative design, artificial intelligence, evolutionary algorithms, deep learning algorithms
Type of Study: Research original/ Regular Article |
Subject:
معماري و شهرسازي
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |