Generative design for excellence (DFX) for a smart built environment: From rule-based imitation to data-driven exploration

Prevailing in the smart built environment (SBE) landscape is the rule-based generative design paradigm, which functions by exploiting decoded or decodable design knowledge. However, conditioning a design on preset rules and past knowledge leads to imitation, not innovation. In contrast, artificial intelligence (AI) can tap into and harness big data (e.g., design drawings in digital format) and its embedded design knowledge, unleashing the imagination of design. This chapter aims to articulate a data-driven generative design paradigm by comparing its strengths and weaknesses with those of the rule-based paradigm and exploring its status quo and development scenarios. We conclude that the knowledge embedded in existing design data offers enormous opportunities for generative design for excellence (DfX) in SBE development; opportunities that AI has potential to enhance towards freeform exploration. However, much research work, e.g., collecting sufficient design data, addressing data bias, and machine-learnable data representation, is needed to improve the SBE development as empowered by the data-driven generative design paradigm.