AI-Driven Digital Twin for Urban Transport Infrastructure Network Operations Optimization

Abstract

Urban transport infrastructure networks play a critical role in enabling efficient mobility and sustainable urban development in smart-city environments. However, many existing digital twin studies in transportation primarily focus on infrastructure monitoring and sensor-based traffic management. At the same time, limited research has explored the integration of artificial intelligence (AI) and digital twin simulation to optimize network-level transport operations. This study aims to develop a conceptual AI-driven digital twin framework for urban transport infrastructure networks to optimize traffic flow, network capacity, and overall mobility efficiency. The research adopts a simulation-based and experimental approach, combining transport network modeling using graph theory with synthetic traffic datasets that represent nodes such as intersections, terminals, and stations, as well as edges representing urban road corridors. The framework incorporates AI-based optimization models, including reinforcement learning and heuristic optimization techniques, to evaluate alternative operational scenarios such as baseline network operations, AI-assisted traffic flow optimization, and adaptive infrastructure management. Simulation results indicate the potential for improved network operational performance under controlled experimental conditions, including reduced congestion levels and more balanced traffic distribution across the network. These findings are limited to simulated environments and do not represent real-world validation. Consequently, the proposed framework provides a scalable, exploratory analytical approach for assessing transport operational strategies and supporting data-driven decision-making in urban transport infrastructure management and smart city mobility planning within simulation-based contexts.