A Dual-Evaluation Generative Design Method for Tension Net Structures
Matt Turlock & Kyle Steinfeld
The nature of design tools is related to the social relationships they serve. This paper speculates on the emergence of a new professional configuration - the synthesis of architect and engineer - and on the nature of new computational tools and methods that will be required to support such a reconfiguration.
Extending previous work that established a framework for the application of Machine Learning (ML) to Generative Architectural Design (GAD), we present here an approach that employs two distinct evaluators: The first stands in for the engineer, and quantifies structural performance; The second stands in for the architect, and assesses candidate designs based on qualitative factors, an evaluation that is made possible by employing a neural net. This new framework is demonstrated through an investigation into tension nets and their structurally derived forms. Since such a tool allows for these evaluators to be employed in combination or in isolation, the resulting solution sets can illuminate both synthetic solutions and each of the two desires independently - a capacity that implies value not only as an optimization tool, but also as a tool for exploration and education.