Automation in architectural design: optimization and decision-making in interference design

The design of architectural projects encompasses a complex array of types, constraints, and contradictions, necessitating a high reliance on the personal experience of architects. Automated generative design aids in exploring various design alternatives, but current research grapples with challenges such as limited user customization, inadequate integration with site environments, and insufficient amalgamation of architectural form and performance optimization. This paper develops an algorithm for early-stage architectural generation, which employs a multi- interference item strategy, serving as a highly customizable design tool. In conjunction with performance optimization based on the NSGA-II algorithm, it introduces a versatile workflow. This paper applied this workflow in a residential building project, whilst incorporating indices such as structural performance, material usage, and site plot ratio as optimization objectives and constraints. The resulting design solutions enhance structural performance while meeting the constraints. The generative design workflow of this study boasts advantages in efficiency, performance, and customization levels. It enables designers to explore the design space in a customized manner, yielding effective solutions within a reasonable timeframe.