AI-Driven Architecture: Automating Space Planning and Building Mass Generation
摘要
Artificial Intelligence (AI) has played a fundamental role in the transformation of various sectors, including architecture and construction. The integration of AI with the Building Information Model (BIM) framework has allowed addressing the specific needs of users effectively in aspects such as floor space planning, building structure generation, and the optimization of material and energy use. This study proposes the development of specific code extensions, adapted to the local regulations of Cuenca, Ecuador, to facilitate the effective application of AI through generative design techniques in the architectural process. These extensions are integrated with the Autodesk Revit architectural design software and its Autodesk DynamoBIM add-on, allowing architects to fully leverage the capabilities of AI while adhering to local regulations. The main objective is to reconcile architectural creativity with regulatory constraints, allowing the generation of aesthetic, optimized, and contextually relevant proposals within the urban framework of Cuenca in much less time than usual. An empirical evaluation was conducted with construction professionals from the “A y D Estudio” architecture firm, assessing the performance and satisfaction with the use of the proposed code extensions for automatic volumetric generation and space planning. The results show significant time savings, high compliance with regulations, and a high level of user satisfaction, indicating the potential for widespread adoption of these tools in the industry.