Les modeles de prevision de deplacements urbains
Institution:
Lyon 2Disciplines:
Directors:
Abstract EN:
Since the end of world war ii, the urban transport sector has benefited from an exceptional grawth which has engendered a great need in studies. The multiplication of such studies has generates numerous technical means and mor specifically models of traffic forecast whether long term ones or short term ones. After a very productive period, with the contribution of a great variety of methods that were developed in the united-states, french forecast came to lack in confidence at the beginning of the 70's. Such a distrust is partly explained by the fact that these models could not anticipate conjunctural crises. In france, such a questioning led to the development of two tendencies. On the hand, the first trend that was more particularly concerned with a practical activity, consisted in improving the existing tools with yet integratind some new developments as those produces, for instance, by the disaggregated approach. On the othe hand, the second trend which emerged from a research activity examined more thoroughly mobility on the basis of a criticism of the usal means. Twenty years later, in spite of improvements due to both trends, there still is an important us of tools that were created in the 50's or 60's. In fact, there has not been any major change as far as urban trip forecast is concerned. Would the french forecast be at a complete standstill? would its development be at a new turn? is this situation typically french? a detailed study covering nearly 40 years of forecast and completed by the analysis of the practical case of lyon gives a certain number of answers. Theses ones will be extended by the presentation of a few prospects of models
Abstract FR:
La croissance exceptionnelle dont a beneficie le secteur des transports urbains depuis la fin de la seconde guerre mondiale, a induit un developpement considerable des besoins d'etudes. La multiplication de ces etudes a donne lieu a l'elaboration d'outils varies, en particulier de modeles destines a prevoir les trafics a un horizon plus ou moins lointoin. Apres une periode faste, qui a vu l'introduction de nombreuses methodes developpees aux etats-unis, la prevision francaise a ete marquee par une crise de confiance, au debut des annees 70. Cette defiance vis a vis des modeles s'explique, en partie, par leur incapacite a anticiper les chocs conjoncturels. Cette remise en cause a debouche au niveau francais, sur le developpement de deux demarches. La premiere, menee en majorite par des praticiens, a consiste a rechercher une amelioration des outils existants, tout en integrant certains nouveaux developpements comme ceux produits par l'approche desagregee. La seconde, en revanche, issue du milieu de la recherche, s'est davantage attachee a approfondir la connaissance des phenomenes de mobilite sur la base d'une critique des instruments conventionnels. Vingt ans plus tard, en depit de certains progres sensibles enregistres par les deux demarches, l'utilisation d'instruments realises dans les annees 50 et 60 reste importante. Aucun bouleversement fondamental en matiere de prevision des deplacements urbains ne s'est veritablement produit. La prevision francaise est-elle dans le marasme? se trouve-t-elle a un nouveau tournant de son developpement? cette situation est-elle typiquement francaise? un bilan detaille de pres de quarante annees de prevision, complete par l'examen d'un cas pratique au plan lyonnais, contribuent a apporter une serie de reponses. Celles-ci seront prolongees par la presentation de quelques perspectives de developpements des modeles.