Empowering Sustainable Smart City Digital Twins with Deep Generative and Foundation Models

This chapter conducts a comprehensive state-of-the-art review to investigate how the integration of deep generative models (DGMs) and foundation models (FMs) into urban digital twins (UDTs) can enhance the environmental planning and computational design of sustainable smart cities by leveraging their synergistic and complementary capabilities. The primary objective is to introduce a novel foundational framework, grounded in theoretical insights and empirical evidence, which outlines the roles, functions, and strategies of DGMs and FMs in current UDT systems. The bibliometric analysis reveals a growing academic focus on the integration of DGMs, FMs, and UDTs, with emerging trends in environmental sustainability. The proposed framework is designed to enhance predictive analytics, enable adaptive and self-learning mechanisms, and improve the synthesis and management of complex, multi-scale urban data. The chapter identifies key challenges, including ethical considerations, environmental costs, and the imperative for interdisciplinary collaboration. Combining evidence synthesis with applied methodologies, the study equips urban planners, designers, and policymakers with both theoretical foundations and actionable strategies for harnessing DGMs and FMs to augment UDT functionalities, particularly in addressing the complex environmental and spatial challenges encountered in contemporary urban contexts. This ensures UDT systems leverage these technologies to shape the landscape of environmentally sustainable smart cities.