Computational-based Generative Design Exploration, Multi-Agent System as an Approach

El-Maghraby, Israa M. and Jahin, Heba and Al-Hagla, Khalid S. M. (2023) Computational-based Generative Design Exploration, Multi-Agent System as an Approach. LET IT GROW, LET US PLAN, LET IT GROW. Nature-based Solutions for Sustainable Resilient Smart Green and Blue Cities. Proceedings of REAL CORP 2023, 28th International Conference on Urban Development, Regional Planning and Information Society. pp. 145-154. ISSN 2521-3938

Text (Computational-based Generative Design Exploration, Multi-Agent System as an Approach) CORP2023_34.pdf - Published Version Download (507kB)

Abstract

Architecture design problems are known for their sequential steps that address a series of several interweaving, competing and/or aligning requirements. Describing them as open-ended, uncertain, solution searching processes makes them complex and ill-structured design problems. One of the methods to deal with nonlinear complex systems is that their components’ properties and features must not be pre-determined and studied linearly in isolation. Instead, it is essential to consider the system as a whole, even if it means considering it generally and roughly, and then allow possible simplifications to occur from the dynamic interactions between components. Computational design methods that encounter distributed computation and artificial intelligence, such as Multi Agent System (MAS), showed promising abilities in addressing complexity and uncertainty faced with architecture design problems, as well as they proved positive effect on expanding architecture design exploration (ADE). This study has an interest in MAS capabilities in creating aesthetically innovative and performable architecture solutions. Therefore, this research intends to investigate the use of MAS in Architecture in the years between 2010 to 2020. It contributes with a detailed examination of research papers to orient future research in the field of MAS. Hence, the applied literature review raises the question of what the proved capacities of MAS are and how future research can challenge it further to widen and develop the use of MAS in ADE and their possible capabilities when addressing building performances such as structural, functional, and environmental. The databases used for selecting these papers are Scopus, Web of Science, SAGE, Science Direct, Google Scholar, Connected Papers, CUMINCAD, IEEXplore, and ACM Digital Library. These studies are organized, analyzed, and compared to pinpoint key innovations in MAS’s variable usage, study its applied methods, interesting results, important sources of data, implementation strategies, and shed the light on the gaps and shortcomings to draw a perspective of MAS in architecture. The examined studies are arranged chronographically. Then, each paper is analyzed and classified according to the aim of the methodology, domain, level, scale of application (experimentation level), model generation and optimization methods. Afterwards, a critical review is proposed.

Item Type:	Article
Uncontrolled Keywords:	Computational Architecture, Design Optimization , Generative, Algorithmic & Evolutionary design, Agent based Modeling and Simulation, Multi-Agent Systems
Subjects:	H Social Sciences > H Social Sciences (General) H Social Sciences > HD Industries. Land use. Labor Q Science > QA Mathematics > QA76 Computer software
Depositing User:	REAL CORP Administrator
Date Deposited:	03 Oct 2023 19:28
Last Modified:	04 Oct 2023 17:31
URI:	http://repository.corp.at/id/eprint/1047

Actions (login required)

View Item