Abstract
The multi-functional space of sports arena is highly related to the long-span structure. To support the integration of these two aspects, design optimization combining parametric modeling, performance simulations, and searching algorithm can be used. However, optimization is powerful in dealing with quantitative performance, but for some soft requirements on buildings, design exploration of geometries based on the judgments of architects is still necessary. Self-organizing map (SOM), as a model-based clustering algorithm, can be used to support this kind of explorations on geometric typology. Nevertheless, it is difficult to ensure the accuracy of clustering, especially for complex parametric models. To support the design exploration on geometry (besides the exploration on quantitative performance supported by optimization) during the conceptual design of sports arenas, this paper proposed a process based on a versatile and flexible parametric model for sports arenas and self-organizing map (SOM). Within this process, to increase the accuracy of SOM clustering, a pre-processing step for the parameters of design alternatives is also proposed. A design of a hypothetic sports arena is used as a case to demonstrate and verify the process.
Original language | English |
---|---|
Title of host publication | Proceedings of the IASS Symposium 2018 |
Subtitle of host publication | Creativity in Structural Design |
Editors | Caitlin Mueller, Sigrid Adriaenssens |
Publisher | IASS |
Number of pages | 8 |
Publication status | Published - 2018 |
Event | IASS 2018: Annual Symposium of the International Association for Shell and Spatial Structures : Creativity in Structural Design - Boston, United States Duration: 16 Jul 2018 → 20 Jul 2018 |
Conference
Conference | IASS 2018: Annual Symposium of the International Association for Shell and Spatial Structures |
---|---|
Country/Territory | United States |
City | Boston |
Period | 16/07/18 → 20/07/18 |
Keywords
- design exploration
- self-organizing map (SOM)
- clustering
- parametric modeling
- multi-objective optimization (MOO)
- multi-functional space of sports arenas
- long-span structure