PLAteformes numeriques INnovantes de Simulation Urbaines pour le Developpement Durable – PLAINSUDD

The operational modelling of personal transport is now relatively well managed and is marked by the availability of a range of models based on sound and standard knowhow. This is not the case, however, when modelling space, whether for the system dedicated to spatial localisation or its ensuing economic attributes such as house prices and social segregation.

This research therefore endeavoured to develop two families of LUTI models in two experimental areas: the Ile-de-France region and Greater Lyon. The two families are, respectively, a dynamic model using the URBANSIM software package developed by Paul Waddell (University of California, Berkeley) and the PIRANDELLO equilibrium model developed by Jean Delons (Vinci Concessions). The methodological organisation entailed getting the two families of model to operate in permanent confrontation with each other in the two areas: URBANSIM was implemented in the Ile-de-France region by IAU-IdF and in Greater Lyon by LET; PIRANDELLO was deployed by Vinci Concessions in both areas. The aim consisted in simulating identical «impacts« in each area, that is to say principally major infrastructure projects in the Ile-de France region and different types of urban toll systems in Greater Lyon.

The first significant advance achieved through this research stemmed from the fact that it was possible to develop these LUTI models to operational level with satisfactory testing of their pertinence. Using the database of IAU-IdF in the case of Paris, and the MOSART numeric platform developed by LET in the case of Lyon, facilitated keeping within the schedule set for project, despite the particularly cumbersome implementation of the localisation models (enhanced by more traditional but highly detailed traffic models involving network digitization).

The most interesting, and in some respects the most original, result was that the effects of strong transport policy initiatives on urban configurations could be simulated over the long term for both the large Parisian projects and the urban toll projects of Lyon.

In the latter case, the simulations presented in the framework of the «Lyon Urban Truck and Bus« (LUTB) cluster must be reutilised in the procedures of several local initiatives.

This experimental research and development project on models was carried out from December 2008 to May 2012 by a UMR of the CNRS, LET (Alain BONNAFOUS, research coordinator), by the companies VINCI Concessions (Jean DELONS and Vincent PIRON) and GEOMOD (Philippe BISCHOFF), and by the IAU îdF (Dany NGUYEN-LUONG). The project received funds of €790,000 from the ANR (Total cost of the project: €1,788,000)

The applications mentioned above correspond in our disciplines to the equivalent of patents in the form of non-disposable models. There is not yet any need to apply for patents or equivalent forms of intellectual protection since specific knowhow is required to implement both these models. However, Vinci Concession has filed the codes of version II of Pirandello with the French National Council of Engineers and Scientists (no. 22 438).

Besides the corresponding models and software packages, this research has already given rise to 5 articles published in chapters of works and in internationally reputed scientific reviews with peer review committees (and 4 are currently awaiting acceptance for publication). More than twenty international communications have been presented on the works and results generated by this research and 5 others have been submitted for the next World Conference on Transport Research (Rio, July 2013).

The increasing concern over the last years about sustainable development has brought forth the usage of quantitative models in the field of transport and spatial planning. This project proposes to improve our understanding of the interaction between transport and land use models and to increase the readability of their results for decision makers at all levels. From a scientific standpoint, the building of large forecast models that span several disciplines of social sciences and natural sciences is a major challenge for the upcoming years. In particular, disciplines that focus on sustainable development must be able to integrate in a consistent way economic models with physical and environmental ones. The interaction between mobility and land use is an ideal example of such integration. So far, there does not exist any framework that integrates state-of-the-art spatial economic models (of housing market, location choice, etc.) and modern transportation models (i.e. time-dependent segmented travel demand and traffic) under the same umbrella and with the support of the latest technology to analyse and disseminate the results. Several research teams have embarked on that exciting challenge in the world: URBANSIM (USA)(Waddell et al.2003), ILUTE (Canada)(Salvini and Miller 2003), ILUMASS (Germany)(Moeckel et al. 2002) among others. The scientific challenge is to bring together the skills from various disciplines: (a) sophisticated traffic models have been mainly developed by engineers without much concern about the socio-economic reasons why travel demand arises in the first place; (b) economists often treat the transportation sector as a black box and tend to prefer the development of aggregate mathematical models; (c) planners and geographers are concerned about spatial development, have the knowledge of the relevant mechanisms and recognize the need for innovative tools but do not have the training to develop complex numerical forecast models; (d) lastly, computer scientists are nowadays the only actors with the appropriate technical training to design such complex models but obviously they lack the theoretical background for modelling cities and mobility. The project plans to build on accumulated experience at several institutions both academic and from the private sector in the field of operational Land Use and Transport Integrated (LUTI) models. The programme of the project includes the ongoing efforts in Paris and Lyon to have operational applications of the OPUS/URBANSIM framework. It also includes the test of an experimental yet promising approach consisting of the PIRANDELLO model. While maintaining this effort, the project also plans to focus on some particular issues regarding the improvement of a) real estate price models; b) location choice models of commercial activities and c) transportation models by using time-dependent approach (METROPOLIS). An important aspect of PLAINSUDD is to dedicate substantial efforts to the development of MOSART which is a web computer platform to disseminate the results of such models. The idea is that the system is going to be able to show the future aspects of the simulated city in ten or twenty years in a manner that is both intuitive and accessible to the decision makers. For this purpose, LET will team with GEOMOD to develop and implement MOSART using the latest web technologies and industry standards such as VISUM transportation modelling software. The prototype of MOSART is meant to be a proof of concept easily transferable to other cities that could be potentially interested in such prediction tool.