Urban road network congestion bottlenecks identification and diffusion analysis

Accurately identifying urban road congestion bottlenecks and analyzing their propagation are crucial for proactive traffic control. This study proposes a network-level method to identify bottlenecks and assess their diffusion risk. A GT-Eclat model integrating topological and time constraints is constructed to identify bottleneck areas. Secondary association mining then identifies congestion diffusion paths, and a Congestion Radiation Index (CRI) is proposed to quantify the diffusion risk. An empirical study in Wuhan, China, demonstrates the method's effectiveness. Results show bottleneck formation stems from the coupling of structural bottlenecks and high traffic demand, with spatiotemporal heterogeneity in congestion intensity. CRI analysis indicates that a bottleneck's diffusivity is less related to its own congestion intensity but more influenced by path structure and traffic flow spatiotemporal characteristics. These findings offer valuable insights for dynamic and differentiated congestion management strategies.