THE STRADA MODEL FOR DYNAMIC ASSIGNMENT

Jean‐Patrick Lebacque, Christine Buisson, J B Lesort, H Mongeot

Intelligent Transportation: Realizing the Future. Abstracts of the Third World Congress on Intelligent Transport SystemsITS America · 1996 · 19 citations · 0 references

Concepts

Abstract

The STRADA model is being developed by ENPC (Ecole Nationale des Ponts-et-Chaussees) and ENTPE-INRETS (Ecole Nationale des Travaux Publics de l'Etat - Institut National de la Recherche sur les Transports et leur Securite). The aim of this model is to provide a computational tool for the simulation of traffic flow under dynamic assignment conditions, and for the implementation and evaluation of various control and guidance strategies implying the use of user-information systems. The STRADA model is the discretized version of a first-order macroscopic traffic flow model that has been extended in a consistent manner both to intersections and to partial flows. The discretization uses a generalization of the Godunov scheme based on the traffic flow supply and demand concepts. As a consequence, a single discretization scheme is applied to both links and intersections, thus providing a simplicity in the description of traffic dynamics that makes the model suitable for the simulation of large networks. The model can be used indistinctly for urban street networks, motorways or corridors. The estimation of the model parameters is straightforward: those are geometrical parameters, or quantities measured as a matter of routine by operators. The traffic flow may be disaggregated according to various criteria such as destinations, ODs, paths, travelers access to information. A prominent feature of the model is its treatment of intersections, which may be modeled with various degrees of refinement, depending on the user's needs and/or capabilities of data collection. Finally, for the treatment of reactive assignment, a simple nested logit user behavioral model has been included into the STRADA model, as well as a new scheme for the estimation of instantaneous and experienced travel times. The method of estimation of these quantities has been designed in order to cope with both interrupted and uninterrupted traffic. Some computational versions of the model have already been realized. The authors shall present in this paper some preliminary results, related to incidents. A framework for the description and analysis of incidents has been developed, consistent with first order macroscopic traffic flow modeling, and has been suitably adapted to STRADA. As an illustration of these ideas, an example of the effect of VMS (variable message signs) on the assignment of traffic in a simple network is described.