Modeling congestion propagation in urban trans- portation networks

In urban transportation networks, research work has been focusing on alleviating the traffic congestion to optimize the network performance, such as increasing the traffic flow and decreasing the user travel time. However, studying the physical properties of traffic propagation and congestion growing has always been a difficult task, especially in a large and complex transportation network. Previous research has seen the main challenges of high complexity of modeling the traffic flow mechanism and the unpredictability of vehicle or user behaviors. The concept of maximum connected component (MCC) is recently used in analyzing the robustness and fragility of a given network structure, such as the catastrophic failure propagation of electrical blackout in a large city network. This paper studies the traffic growing and propagation in a large network based on the concept of MCC, without modeling the traffic flow mechanics. Firstly, a congestion propagation model is developed and a metric evaluating the congestion was introduced. Secondly, simulations with different parameters are developed based on the model. The results are compared with real traffic data from San Francisco network and the model is validated. It mainly shows the strong dependency of congestion among links. Finally, we show that the propagation effect is linked to the scaling law, which is a key property of the urban transportation networks.