Characterizing Expert Behavior During Interactive Parametric Building Design

Parametric design strategies allow experts in the Architecture-Engineering-Construction industry to rapidly consider many potential options in early building design. Optimization techniques applied to parametric models can systematically find the best options in terms of quantitative design goals such as energy use or structural efficiency. This project will investigate how optimization techniques can best assist designers in managing qualitative and quantitative goals simultaneously. In a broader contex, the project will study how emerging human-computer collaborative tools influence design choices, how cognitive processes relate to the outcomes they produce, and especially how they are contextualized in architectural engineering. The project will characterize the behavior of expert designers as they employ optimization-based parametric design methods to formulate and explore design options in early-stage design. The project includes educational and outreach activities such as industry workshops to retrain established, traditionally analog-trained engineers in new digital design techniques and software. These activities will help to prepare a workforce to be computationally agile in their careers, with the ability to use digital tools to design buildings that are more energy-efficient, safe, durable and sustainable.

This project will employ an empirical multiple methods research design to investigate experienced practitioners' design strategies using eye tracking and observational data. The main research question under study is: 'What patterns of design behaviors do architectural engineers employ while constructing and exploring a parametric model using optimization-based tools?' Expert building designers will be recruited from architecture and engineering design firms and will participate in a design task employing data-informed parametric modeling. Their behaviors will be captured using time-resolved recorded data of the authentic design task from the design software and eyetracking hardware. The multiple streams of data will be analyzed using machine learning capabilities such as Hidden Markov Modelling and other statistical methods to characterize behaviors. The project will provide a foundation for evidence-based approaches to computational design tool development and will contribute to design theory in digital interfaces, optimization, parametric modeling, analysis of multiple real-time data streams. A related objective is to improve education of established design professionals in the use of emerging visual programming interfaces, parametric design, and optimization concepts.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.