Clustering route choices methodology for network performance analysis - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue Transportmetrica B-Transport Dynamics Année : 2017

Clustering route choices methodology for network performance analysis

Résumé

The Network Macroscopic Fundamental Diagram (NMFD) provides a quick overlook of a network performance. Recent studies show that inhomogeneous density spatial distributions significantly reduce the network efficiency. However, few studies investigate the triggers of such distributions. In this paper, we investigate how route choices influence the level of performance of a network and we propose a complete methodology to analyse such a relationship. To simplify the exploration, route choice alternatives are clustered in homogeneous groups with respect to the percentage of overlapping and a large set of flow distribution among routes is tested. The investigations are conducted using both analytical formulations and simulation techniques based on a mesoscopic traffic flow simulator. Analytical formulations define the boundary between fluid and congested network loading. The sensitivity to route choices is then assessed. Then, route choice effects on the network level of performance is quantified using the NMFD. The different evaluated scenarios confirm that the network performance is highly affected by route choice, both by paths' selection and flow distribution. Moreover, the study of hysteresis loops emerging in the NMFD shows that route choice impacts significantly the spatial distribution of congestion.
Fichier non déposé

Dates et versions

hal-01301683 , version 1 (12-04-2016)

Identifiants

Citer

Céline Parzani, Ludovic Leclercq, Nazih Benoumechiara, Daniel Villegas. Clustering route choices methodology for network performance analysis. Transportmetrica B-Transport Dynamics, 2017, Special Issue on Traffic Flow Theory, 5 (Issue 2), pp. 195-214. ⟨10.1080/21680566.2015.1135765⟩. ⟨hal-01301683⟩
142 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More