Design Agency: Prototyping Multi-Agent System Simulation for Design Search and Exploration

摘要

This work presents prototypes of multi-agent system simulation for design search and exploration. We describe an experimental approach in part based on a previously established multidisciplinary design optimization (MDO) framework. Here the work further explores the potential impact of MDO in conjunction with multi-agent systems on the early stages of design. Specifically, this paper addresses the potential of introducing agent based computing techniques into the multidisciplinary architectural design optimization and search workflow to tackle geometrically complex design problems and to facilitate early stage design exploration. To address these interests a series of prototyped workflows were studied inclusive of environmental performance and structural performance metrics and benchmarks. This paper presents a novel methodology for using simulation data in conjunction with multi-agent systems as a way for re-informing form and enhancing performance in a generative design environment. The methodology is based on the use of swarm algorithms and their integration with data generated by simulation software. The interaction between these two domains, the simulation data and swarm algorithms, generates the final outputs as a modified geometry that is then evaluated by comparison for enhanced design performance.