Abstract : Network flow computing based on macroscopic traffic flow models for large and dense networks involves a large number of parameters and variables, and significant computational efforts. We aim at reducing these and introduce a modelling framework at two-dimensional scale in order to model traffic flow of transportation systems of large surface networks. We present a network flow pattern corresponding to network flows modelling with a few network sensors of traffic count locations. We manage and evaluate traffics on wide and dense networks with a minimum of available measurements and data, through modelling of global behaviours based on local behaviours. We find that the traffic at this scale is governed by multidimensional hyperbolic conservations laws. Godunov-type method has been proposed to compute the network flow flux across computational domains.
https://hal.archives-ouvertes.fr/hal-01559819
Contributeur : Kwami Sossoe
<>
Soumis le : vendredi 14 juillet 2017 - 03:01:17
Dernière modification le : mercredi 11 avril 2018 - 12:10:02
Document(s) archivé(s) le : vendredi 26 janvier 2018 - 19:59:00
