From Cellular Automata to AI and Fabrication

Abstract

This paper presents research on workflows for architectural design through the integration of Cellular Automata (CA), Artificial Intelligence (AI), and Digital Fabrication. This study aims to explore how these technologies can optimise both structure and aesthetics in architectural design, providing an efficient and innovative framework. The methodology combines the generation of complex patterns with CA, the use of AI diffusion models for geometry optimisation, and the digital translation of these designs using Autodesk Fusion, enabling comprehensive project management and facilitating interdisciplinary collaboration. The methodology focuses on generating Cellular Automata (CA) patterns using a combination of bespoke and existing algorithms (e.g., Langton's Ant). CA patterns are developed within the Processing environment, leveraging local rules to create complex configurations. AI models, specifically diffusion models, are employed to iterate and optimise these geometries, facilitating continuous refinement of architectural designs and ensuring precise adaptation of the CA-generated patterns. This work develops a workflow aligned with the adoption and transformation of technologies in the Global South, particularly by integrating AI into the creative process. It opens pathways for adapting these methodologies to resource-constrained contexts, proposing a comprehensive framework that combines innovation and efficiency. The study significantly contributes to the future implementation of these technologies, promoting their accessibility and adaptability.