Exploring Alternatives with Design Analytics Interfaces: A Human-Centered Approach for Integrating Generative Design and Performance Assessment with Machine Learning-Based Surrogate Modelling

Designing built environments is a complex process that involves generating and evaluating alternative solutions using computational tools. In the Architecture, Engineering, and Construction (AEC) industry, designers use generative design methods to explore a larger number of alternatives and ML-driven rapid performance prediction to identify potential issues early on. However, separating these tasks hampers creative flow in decision-making. This project aims to bridge the gap between generative design and performance assessment with ML-based surrogate modelling techniques by introducing novel Design Analytics tools for AEC projects. We aim to enable designers to search for designs satisfying design criteria for better building performance, enhanced energy efficiency, and reduced environmental impact, ultimately advancing the sustainability and resilience of built environments. Our approach will incorporate interactive data visualizations to support decision-making and enable efficient design space exploration.

Faculty Supervisor:

Halil Erhan

Student:

Partner:

Perkins+Will Canada Architects Co.

Discipline:

Engineering

Sector:

Construction; Information and Communications Technology; Technology

University:

Simon Fraser University

Program: