AbstractsComputer Science

Swarm Intelligence: Defining Agent-based Design Methodology

by Mohd Syafiq Azmy




Institution: Savannah College of Art and Design
Department:
Year: 2016
Keywords: Thesis (M.Arch.)  – Architecture; Savannah College of Art and Design  – Department of Architecture
Posted: 02/05/2017
Record ID: 2078332
Full text PDF: http://ecollections.scad.edu/iii/cpro/DigitalItemViewPage.external?sp=1003467


Abstract

'The use of swarm intelligence nowadays has expanded not only to the field of science, but also can be seen in architectural explorations especially in the design process. Such explorations in architecture need to be defined in terms of how swarm intelligence can be segue into architectural typologies that focus on form, structure and site. This paper describes how agent-based design methodology (swarm intelligence) act as a vehicle in the design process in defining architectural pavilion typologies. The study begins with addressing comparisons between analog, digital analog, digital parametric and agent-based design methodology. Next, the interaction between agent-based methodology with uncontrolled swarm experimentis studied. The processes developed during the first experiment is use to look at the possibilities of swarm parameters that can be use for controlled swarm for form, structure, and site simulations. In general the paper study highlights the definition of swarm, theory of swarm intelligence, and presents an example of case studies on how swarm can generate specific architectural typologies. This allows the development of agent-based methodology workflow and programming algorithms that are used in the experiment simulations. The experiment simulations uncovered the creation of specific architectural typologies that do not belong into the realms of abstract computation but necessary for the role of designer to engaged in the experiments in order to control any number of parameters, and to provide the desireable architectural design intent that meet with certain set of architecture criteria such as structure opening size and thickness.' Keywords: swarm intelligence, agent-based design, computational design, grasshopper3D, pavilion, self-organisation, simulation, digital design, swarm methodology Advisors/Committee Members: CHAIR: Singeisen, Scott R., Ngo, Huy Sinh, Riether, Gernot.