Jiping, Xing (Nanjing Forestry University), Jia, Zhou (Southeast University)

Leveraging LLMs for Resilience Evolution-Based Traffic Congestion Regulation: Review and Perspectives

Scheduled for presentation during the Invited Session "S18b-Innovative Applications of LLM in Multimodal Transportation Systems" (TH-EA-T18), Thursday, November 20, 2025, 13:50−14:10, Southport 3

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 18, 2025

Abstract

Urban traffic congestion manifests in diverse forms, with its evolution and propagation patterns being influenced by multifactorial determinants. The formulation of efficient traffic control measures for congestion mitigation is frequently constrained by dynamic variations in road network traffic flow percolation states, particularly evident in abnormal congestion scenarios where temporal effectiveness constraints emerge. Current research methodologies pertaining to road network resilience evolution metrics enable the implementation of differentiated regulation strategies through phase-specific congestion characterization. The advent of large language models (LLMs), empowered by robust algorithms and massive datasets, demonstrates distinctive advantages in providing technical support for future high-efficiency rapid processing of complex traffic percolation phenomena and congestion regulation. This study systematically synthesizes existing knowledge through two primary dimensions: (1) the application of resilience metrics for alleviating network-wide congestion, and (2) prospective enhancements of congestion management efficiency through LLM integration. Furthermore, we delineate promising directions for subsequent investigations, emphasizing the need for interdisciplinary convergence in intelligent transportation system optimization.