A Multi-Step Approach for the Global Sensitivity Analysis of Complex Traffic Simulation Models: Application to the MITSIM Model

In this work a multi-step approach for model sensitivity analysis has been discussed and applied to the MITSIM traffic simulation model. Throughout the paper it is argued that the application of sensitivity analysis is crucial for a true comprehension and the correct use of traffic simulation models, but it is also acknowledged that the main obstacle towards an extensive use of the most sophisticated techniques is jeopardized but the high number of model runs usually required, especially in the case of models with a high number of model parameters. For this reason the authors have tested the possibility to perform a number of preliminary analyses by grouping model parameters on the basis of possible common features (in particular grouping all the parameters pertaining to the same sub-models) and, then, to use sensitivity analysis to discover which groups of parameters accounts for the highest share of the outputs’ variance. At the end of these preliminary steps a final sensitivity analysis on the parameters pertaining to the most influential groups can be performed to individuate the most important among them. The proposed methodology has been applied to the MITSIM model (101 model parameters) and has allowed uncovering the role played by the different parameters and by the model stochasticity with 80% fewer model evaluations.

  • Supplemental Notes:
    • This paper was sponsored by TRB committee AHB45 Traffic Flow Theory and Characteristics.
  • Corporate Authors:

    Transportation Research Board

    500 Fifth Street, NW
    Washington, DC  United States  20001
  • Authors:
    • Ciuffo, Biagio
    • Azevedo, Carlos Lima
  • Conference:
  • Date: 2013


  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; References; Tables;
  • Pagination: 22p
  • Monograph Title: TRB 92nd Annual Meeting Compendium of Papers

Subject/Index Terms

Filing Info

  • Accession Number: 01477173
  • Record Type: Publication
  • Report/Paper Numbers: 13-3726
  • Files: PRP, TRIS, TRB, ATRI
  • Created Date: Apr 1 2013 10:50AM