Digital fabrication technologies greatly enhance and extend manufacturing possibilities. However, the ability to fully exploit these new methods and create complex architectural structures with performance-driven properties are still relatively limited. In this paper by Daniel Richards and Martyn Amos, authors argue that entirely new computational approaches are needed, using scalable generative encodings and advanced bio-inspired form finding processes.

The paper presents a novel generative model that can create functional and expressive geometries by evolving volumetric gradient patterns. Using three case studies, authors demonstrate the key advantages of their approach. They demonstrate, using simulation followed by physical fabrication, that their approach is useful for exploring complex, yet buildable geometries in early stage design. Their new method is therefore suitable for performance-driven form finding tasks such as structural optimization, and holds vast potential for designing exotic multi-material and functionally graded materials in future applications.