Empirical analysis of congestion spreading in Seoul traffic network

TL;DR

This paper empirically investigates how traffic congestion spreads in Seoul's urban network, revealing different spreading patterns during morning and evening rush hours and highlighting the roles of diffusion and stacking processes.

Contribution

It introduces a congestion identification method applicable to various traffic flow types and uncovers distinct congestion spreading structures in Seoul.

Findings

Congestion spreads as a tree-like structure in the morning.

Loop structures dominate congestion during the evening.

Diffusion and stacking are key in congestion formation.

Abstract

Understanding how local traffic congestion spreads in urban traffic networks is fundamental to solving congestion problems in cities. In this work, by analyzing the high resolution data of traffic velocity in Seoul, we empirically investigate the spreading patterns and cluster formation of traffic congestion in a real-world urban traffic network. To do this, we propose a congestion identification method suitable for various types of interacting traffic flows in urban traffic networks. Our method reveals that congestion spreading in Seoul may be characterized by a tree-like structure during the morning rush hour but a more persistent loop structure during the evening rush hour. Our findings suggest that diffusion and stacking processes of local congestion play a major role in the formation of urban traffic congestion.