Title: Global sensitivity analysis techniques to simplify the calibration of traffic simulation models. Methodology and application to the IDM car-following model
Authors: CIUFFO BIAGIOPUNZO VINCENZOMONTANINO Marcello
Citation: IET INTELLIGENT TRANSPORT SYSTEMS vol. 8 no. 5 p. 479 – 489
Publisher: INST ENGINEERING TECHNOLOGY-IET
Publication Year: 2014
JRC N°: JRC85031
ISSN: 1751-956X
URI: http://digital-library.theiet.org/content/journals/10.1049/iet-its.2013.0064;jsessionid=4r494c9iao7cf.x-iet-live-01
http://publications.jrc.ec.europa.eu/repository/handle/JRC85031
DOI: 10.1049/iet-its.2013.0064
Type: Articles in periodicals and books
Abstract: As models are simplifications of reality, the management of the uncertainty arising along the whole modelling process is a crucial and delicate operation, which primarily affects the credibility of model results. In the field of traffic simulation to tackle this issue it is common practice to include the model uncertainty alongside the uncertainty in the parametric inputs. However, reducing the uncertainty in the modelling process through the indirect estimation of the model parameters is far from being simple. In this picture a key role can be played by model sensitivity analysis. In the present work, in particular, the role of sensitivity analysis in the management of modelling uncertainties is firstly illustrated. Then, one of the most advanced techniques to perform sensitivity analysis is explained and applied to identify, in the specific context of application, which of the input factors of two car-following models can be fixed without appreciably affecting a specific output of interest. Results confirmed the relevance of sensitivity analysis in driving analysts’ activities for models’ comprehension, calibration and validation, namely, for their appropriate use.
JRC Directorate:Energy, Transport and Climate

Files in This Item:
There are no files associated with this item.


Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.