A Mesoscopic Large Scale Traffic Model of Place de la Bastille, Paris



A Mesoscopic Large Scale Traffic Model of Place de la Bastille, Paris

Authors

Joan Roca, TSS - Transport Simulation Systems Sarl., Matthieu Jacquart, SYSTRA, Aurore Remy, TSS - Transport Simulation Systems Sarl.

Description

The paper will categorize and explain the different types of simulation model currently being used in Paris, from macro, to micro, and focusing in particular on the meso level modelling of Place de la Bastille..

Abstract

The City of Paris has a range of important development plans for some of its central squares. Some have already been restructured such as the well-known Place de la République, next will be Place de la Bastille. In this case the City of Paris, which has been using for some years a multilevel macro-micro Aimsun traffic simulation package, has decided to develop a large scale mesoscopic model in order to capture in a single model the best of the macroscopic and the microscopic outputs in one go. This paper will detail why the mesoscopic model is the most suitable level of modelling for this study area (12.81 km2), detail the methodology of how the model has been generated and validated, and show the outcomes and applications of such type of models.

The macro model is used to estimate long-term planning issues and is characterized by modelling demand as a constant stream of flow even if greater than supply. The main weakness of this model is that it does not analyze the dynamic aspects of traffic, such as congestion and especially its spread in time. The microscopic model is characterized by an individual representation of vehicles whose behavior is reassessed at regular intervals. This type of model however, requires a level of fairly detailed calibration and significant computing time. The mesoscopic level enables the user to meet various needs, such as representation of traffic redistribution on the whole area of influence and the assessment and evolution of congestion based on the dynamic aspects of supply and demand. It also allows the modelling of the traffic distribution between competitive routes using a DUE algorithm (Dynamic User Equilibrium) with acceptable computation times and simplified calibration.

This latter model covers two three hour peak demands profiled every fifteen minutes for three vehicle types. 315 control plan diagrams and all PT lines were imported together with the detail of the road infrastructure. Aimsun offers the possibility to jump from one modelling level to the next so all these inputs can be used at the micro or macro level (e.g signal diagrams are used as turning penalties for the macroscopic level). Static equilibrium paths are used as initial solution for the DUE and the definition of dynamic parameters such as reaction time, look-ahead distances and jam densities are determined at the global and local level. The validation of the model is carried out evaluating the GEH indicator with simulated and observed counts, and travel time from floating car data of certain itineraries located in the study area.

The papers’ main aim is to categorize and explain the different types of applications these types of models can have, detail the best methodology in each case and represent different cases with specific examples. Two major types of scenarios exist for which the methodology varies:
a) The operational Scenario: This type of scenario aims to model the impacts of specific events or incidents that take place during a short-enough period of time that users do not adjust to it. In this case, the route choice pattern determined in the base scenario is reproduced. Moreover, the temporary alterations due to these incidents/events (e.g. accident, demonstration) is added to the model. It will allow the modeler to analyze the impacts of these incidents on traffic and also to design and test mitigation plans or evaluate the effects of Intelligent Transport Systems.
b) The planning Scenario: The second type of scenario aims to analyze the behavior of users in response to permanent changes to the model, such as infrastructural changes or development of new attraction or generation zones. In this case, the model is used as a planning tool to examine traffic redistribution and the evolution of congestion obtained from a new dynamic user equilibrium process, representing drivers getting used to these permanent changes.

The scalability of the model also allows its extension to all Paris intra-muros which is actually a task under development. The main goal is that the City owns a model that can evaluate the impacts of changes in supply and/or demand for the entire city. This model could also evolve to be connected to the city's control center in order to monitor traffic in real time and use the simulation and traffic management models as a real-time decision making tool in case of non-recurrent events.

Publisher

Association for European Transport