Modele d'equilibre dans les reseaux de transport en commun: Le cas des capacites explicites des services (French text).

摘要

In this thesis we study the static transit assignment problem in general transit networks, which are subject to congestion and have strict limits on the capacity of transit line segments. The problem consists in determining the route choice of transit users given that they cannot access transit line segments, which are at capacity.; The choices made by the passengers depend on three factors: the in-vehicle travel time, the waiting time at the stops and the probability to board a bus of a given line which belongs to a set of attractive lines. When a transit assignment model considers explicit capacity constraints and the above-mentioned factors it is referred to as a “fully congested model”. This is the case under study in this work.; A new definition, that of node-strategy permits a new formulation of this problem. We present a new characterization of the route choice equilibria based exclusively on the arc-flows, which leads to a new model formulation. We obtain a minimization problem of a gap function; the global optima (with zero value of the objective) correspond to transit network equilibria. A heuristic method is proposed to obtain a numerical solution of this problem. The gap function allows the evaluation of the variation of the solutions obtained with this heuristic. The results are satisfactory in the sense that they provide acceptable values of the gap function in a reasonable number of iterations. The empirical convergence of the algorithm is satisfactory.; Another contribution of this thesis is the development of waiting time models at transit stops. These models allow the estimation of the average waiting times of passengers that wait at stops in order to board transit vehicles. Theoretical results demonstrate that the waiting time strongly depends on the other flows in the network and not only on the boarding flow. The best of these stop models incorporates the effect of alighting passengers, the on board passengers and the passengers waiting to board given a limited capacity of the transit vehicle. Moreover, we show that the models that consider exclusively the mean residual capacity of the buses can substantially underestimate the waiting time.