Holistic Design Explorations of Building Envelopes Supported by Machine Learning

Abstract

The design of building envelopes requires a negotiation between qualitative and quantitative aspects belonging to different disciplines, such as architecture, structural design, and building physics. In contrast to hierarchical linear approaches in which various design aspects are considered and conceived sequentially, holistic frameworks allow such aspects to be taken into consideration simultaneously. However, these multi-disciplinary approaches often lead to the formulation of complex high-dimensional design spaces of solutions that are generally not easy to handle manually. Computational optimisation techniques may offer a solution to this problem; however, they mainly focus on quantitative aspects, not always guaranteeing the flexibility and interactive responsiveness designers need in the early design stage. The use of intuitive geometry-based generative tools, in combination with machine learning algorithms, is a way to overcome the issues that arise when dealing with multi-dimensional design spaces without necessarily replacing the designer with the machine. The presented research follows a human-centred design framework in which the machine assists the human designer in generating, evaluating, and clustering large sets of design options. Through a case study, this paper suggests ways of making use of interactive tools that do not overlook the performance criteria or personal preferences.