Agent Based Modeling in Architectural Design
Beginning with the use of Cellular Automata, methods based on individual, autonomous, and self-organizing â€œagentsâ€ have led to a paradigm shift in the approach towards the modeling of complex systems. Agent-based modeling systems (ABMS), for instance, do not yield an overarching abstracted system definition, but merely draw their explorative potential from implementing models of interactivity among singular system participants. Numerous examples of player-based Game Theory have been introduced and studied excessively in the fields of social science and economics.
In architectural or building-related applications, agent-based models help us understand and predict a system's behavior as it evolves on a larger scale through non-linear or quasi-chaotic interdependencies of system agents. Evacuation simulations or the planning of highly integrated network infrastructures have proven to be beneficial applications of ABMS. The outstanding quality of "agents" is their characteristic as self-contained learning units with the ability to permanently reposition themselves within their system environment, while adhering to a set of flexible behavioral rules. The Euclidean interpretation of "repositioning" as a phenomenon of spatial activity opens new perceptions of ABMS as generative architectural tools that integrate complex system constraints into real-time computational "Voxel"-events.
In this seminar, we will explore the fundamental differences between an integrative design approach based on ABMS and the traditional, descriptive, representational design tools of architecture. We will learn relevant programming skills in a compact block workshop, and subsequently study a rule-based approach of spatial self-organization through the creation of individual design tools.
Programming skills are not required.