A Multi-class Multi-mode Variable Demand Network Equilibrium Model with Hierarchical Legit Structures
FLORIAN M and WU J J, University of Montreal and HE S G, INRO Solutions, Canada
Over the past twenty years, the models used in transportation planning for congested urban networks received a lot of attention from researchers and practitioners. One of the principal issues studied and debated is that of the consistency of the models us
Over the past twenty years, the models used in transportation planning for congested urban networks received a lot of attention from researchers and practitioners. One of the principal issues studied and debated is that of the consistency of the models used for predicting the demand for travel and the network models used to determine the levels of service. One seeks to ensure that the levels of service which are used to compute the demand would be reproduced if that demand were used again to compute the levels of service. The keywords that we find in the literature regarding these models are "integrated", "combined", "simultaneous", "feedback" and "variable demand". The recent San Francisco Bay Area lawsuit (see Garrett and Wachs, 1996, p. 199) shows that the importance of this issue has social and political impacts.
The model presented in this paper is inspired from the work carried out for the ESTRAUS project in Santiago, Chile (ESTRAUS, 1985-1998) and is a complex multi-class multi-mode variable demand network equilibrium model with hierarchical structures which, until now, was not formulated in an integrated way as a variational inequality formulation. This formulation permits the dgorous mathematical analysis of the model structure and the development of an efficient solution algorithm.
The remainder of the paper is organized as follows. In the next section, we provide a literature review of the related problems. In section 3, we introduce the notations and develop the variational inequality formulation. Section 4 provides the analysis of the mathematical structure of the model by using the corresponding KKT conditions, and the solution algorithm is given in section 5. In section 6, we provide computational results, and some conclusions are presented in the last section.
Association for European Transport