Creation Through Variation: Cross-Disciplinary Practices in Architectural Design

This presentation covers aspects from my interdisciplinary practice-based research on automating design processes in the context of modular stadia design. The practice side of my research at the Royal College of Art is based entirely on digital workflows and I present my take on digital representation and simulation of built and conceptual designs and their subsequent analysis with machine learning. The presentation focuses on how we can utilise one of the most powerful aspects of computational design – the ability to create large sets of variations within parametrically defined design space and the selection of best performing solution. I present the methods which I used to create the framework of variation through identification of quantifiable design characteristics, the construction of a parametric model through the creation of relationships between those properties, the development of design space, and the consequent reduction of those solutions to identify the design that is to be created.