Resumen
Nowadays, urban population is witnessing a rapid growth and henceforth its mobility. City infrastructures and economic resources may not follow at the same rate as the increasing mobility. So often, projected increase in transport needs goes beyond projected expansion of transit network capacity. This asymmetry between transportation supply and demand is unmistakable: congestion, unpleasant travelling conditions and other phenomena that comes with are already witnessed in the public transit system. Public transport performance is constrained not only by its availability but also by its capacity. Actually, the capacity of a transit line is defined by the operating frequency as well as the physical capacity of each vehicle. The relationship between loaded demand and capacity contributes to the setting of comfort levels in particular and the quality of service in general. To simulate these phenomena in an assignment model describing the users? route and mode choices, the transportation supply should be subject to several constraints: capacity of the vehicles (sitting and standing places), boarding and alighting movements, and the lines and network load (operating frequency of each line). Till recently, in most case studies including new transportation projects, crowding was not taken into account in transport modelling, while congestion should be integrated in simulation, particularly in the objective of effective traffic management. In this paper the studied approach includes three parts. The first part consists of a literature review of French and International existing researches on the influence of crowding in public transport and how it is modelled. The second part describes the adopted crowd-methodologies in our cases studies. The third part outlines the results of three case studies. A conclusion provides a synthesis of the compared results and recommendations. The results are based upon three selected projects conducted by SYSTRA: (1) The extension of the current Metro network and restructuring the bus network in Baku (Azerbaijan); (2) The feasibility study of the Guayaquil (Ecuador) cable line taking into account the existing transit networks (BRT and Bus); (3) The assignment model of Saint-Etienne's (France) public transit (TER ?regional train-, Tram, Bus and Coach) with accurate simulation of passenger trips and capacity constraints. The crowd-model aims at achieving iteratively a balance between loaded demand and capacity. Three approaches are tested for the three projects. First, by adjusting the travel time alone, then by revising the waiting time and finally by combining the two previous approaches. This benchmark highlights indeed some phenomena, such as traffic breakdowns for instance, related to congestion in comparison to a reference scenario without capacity constraints. However, the results show complex spatial and temporal behaviors making their interpretation sensitive.