Robustness Analysis for Road Networks. A Framework with Combined DTA Models

This book describes how network robustness is an important topic for road networks, but there has not been paid a lot of attention towards it. Compared with the concept of reliability, which focuses on analyzing the probability of a road network performing to its proposed service level and adequately taking into account the uncertainties of the circumstances, network robustness problems emphasize particularly on the ability of a road network functioning properly facing unpredictable and exceptional incidents. The robustness studies in other network domain and traditional robustness studies for road networks only deal with the basic connecting features of the network with graph theory, which are basically defined as the connectivity robustness. However, the performance of a road network, such as the speed and throughput that can be felt and observed by travelers and road authorities are also essential aspects for the evaluation of the service level of road systems. This thesis deals with the design and development of a systematic framework that enables the complete network robustness analysis comparing the network performance between daily traffic conditions and incident conditions. In order to accurately achieve the network performance and its changes along with time and space, interactions between the travelers (choice) behaviors and the network performance are important factors to be taken into account. To correctly model this phenomenon as well as the performance of road networks, dynamic traffic assignment (DTA) models play core roles in representing the (path) choice behaviors of travelers and simulating the propagation of traffic flows over the network. Basically, DTA models can be distinguished into two categories: user equilibrium (UE) assignment models and en-route assignment models, according to their basic assumptions about the network status and travelers behavior. In the existing robustness studies for road networks that involve the analysis of the network performance, UE assignment models are implemented in the majority. However, according to the definition and requirements of robustness analysis, the straightforward UE approach is insufficient to analyze the network performance when unpredictable and exceptional incidents occur. Thus, in the proposed framework, both DTA approaches are integrated with the aim to represent the road network performance of daily normal condition and incident conditions. To represent the daily normal condition, (stochastic) UE assignment models are considered as the most suitable and valid tools. For incident conditions, if the occurrence of the incident is unpredicted and exceptional, such as the traffic accidents on the road, en-route assignment models are believed more appropriate than the UE approach to represent them because such approaches are able to model travelers' path updating behavior on their way. Furthermore, by testing different DTA models with a common network, remarkable differences among the results demonstrate that different models have their own specifications and specialties so that it is difficult to compare them. It also means that it is also impossible to combine different models in our framework due to the compatibility problems.

  • Availability:
  • Corporate Authors:

    Netherlands TRAIL Research School

    P.O. Box 5017
    Delft,   Netherlands  2600 GA
  • Authors:
    • Li, Minwei
  • Publication Date: 2008


  • English

Media Info

  • Media Type: Print
  • Features: Appendices; Figures; Maps; References; Tables;
  • Pagination: 172p

Subject/Index Terms

Filing Info

  • Accession Number: 01141035
  • Record Type: Publication
  • ISBN: 9785055841097
  • Report/Paper Numbers: T2008/14
  • Files: TRIS
  • Created Date: Sep 29 2009 12:22PM