XVIII. MİMARLIKTA SAYISAL TASARIM ULUSAL SEMPOZYUMU
An Artificial Intelligence-Assisted Architectural Design Methodology: Integration of Analog Methods, Computational Design and Diffusion Models
Yayıncı:
Balıkesir Üniversitesi Yayınları
Design is inherently complex, often involving ill-defined problems with unclear conditions and unpredictable solutions, especially in early stages. These problem types evolve into well-defined problems as the design process advances. the design process can be divided into four stages as problem structuring, preliminary design, development, and detailing. Each requires different types of information and design media. The rise of Artificial Intelligence , particularly Diffusion Models (DMs), is reshaping the architectural design. Although a comprehensive methodology encompassing their full potential remains underdeveloped, these models, which can generate text and images from prompts, are now being integrated into various phases of the design process. This paper explores how AI and DMs can be seamlessly incorporated into architectural design to enhance creativity and efficiency. It proposes a methodology that integrates traditional analog and computational design methods with DMs throughout the design stages. Initially, DMs stimulate creative ideas by generating images, aiding in the preliminary design phase. This is followed by the generation of alternative designs, where DMs help refine massing models and material selections without altering core decisions. Subsequently, these 2D images are transformed into 3D models, which are further enhanced through algorithmic design tools. Lastly, realistic final representations were made with the help of DMs. The methodology emphasizes a human-centered approach, where AI supports rather than replaces the designer, fostering a reflective practice. A case study evaluating this integrated framework shows that DMs facilitate a dynamic dialogue between the designer and the tool, expanding the search space and enabling more informed decisionmaking throughout the design process. This flexibility not only improves current practices but also sets the stage for future advancements in integrating AI into architectural design, promising new strategies for the field.
