Multi-modal Assignment Techniques of European Freight Transport Chains
MEIJEREN J VAN and CHEN M, NEA Transport Research and Training, The Netherlands
This paper discusses multimodal assignment techniques of European freight transport chains. Based on route characteristics such as distance, tariffs, time, transhipment tariffs and transhipment time the model determines the probabilities of alternative ro
This paper discusses multimodal assignment techniques of European freight transport chains. Based on route characteristics such as distance, tariffs, time, transhipment tariffs and transhipment time the model determines the probabilities of alternative routes between origin and destination. These routes can differ due to competing modes, competing transhipment sites and different link choice options.
The main objective of this research project was to construct and develop a multimodal assignment model explaining freight transport chain choices on a European scale that can be used to estimate freight transport chains where statistics are not available (anymore). This application of the model becomes more and more important because of the deterioration or the disappearance of European freight statistics. Another application of the model is a scenario analysis, for example the results of a fuel price increase or the effect of new infrastructure.
In the second section of this paper the background of this research project is discussed in more detail. After this, in the third section two well-known assignment models are described in general: a Iogit model and a probit model. These models have been implemented and the results of the calibration of both models on a number of datasets from the NEAC inter-regional European transport chain database are given. The probit model is approximated by a simulation model. On the basis of these results the models are compared. The models described in section three show some shortcomings and a number of extensions to the models have been implemented that are described in section four. The following extensions have been implemented:
* Box-Cox transformations of the explaining variables
* the use of stochastic value of time instead of a static value of time
* the implementation of a nested Iogit model
To show the usefulness of such models, in section five a number of applications are presented. Finally in section six a summary, conclusions and recommendations are given.
Association for European Transport