A Multi-agent Model of Land Use and Transport to Study the Dynamics of Urban Mobility ?MobiSim SMA?
Y Martineau, ATN sa, FR
The modelling shell MobiSim SMA integrates the processes that structure urban areas to understand and manage urban mobility in European cities for the next 20 years. The underlying multi-agent model is based on household and enterprise behaviours.
Urban Mobility is a key issue of sustainable development. The modelling shell MobiSim SMA aims to integrate the diverse processes that structure urban areas in order to study, understand and manage the dynamics of urban mobility in French and European cities for the next twenty years. MobiSim SMA is a multi-agent model mainly based on the description of household and enterprise behaviours.
Several agent types are distinguished: dynamic agents (households and enterprises), static agents (road infrastructure and public transportation networks), spatial agents (urban zones) and global agents (institutional managers, lobbying groups, transport associations). The dynamic agents evolve at a yearly time-step in size, number, location in urban area and activities. The model takes into account the demographic events (birth, death, marriage and divorce, child emancipation), the economic transitions of individuals (activity/inactivity, work/unemployment), the residential choices of households, macro-economic scenarios (unemployment rate, long-term trends in education, motherhood and cohabitation behaviour), the location strategy of enterprises, and housing and labour markets. The static agents characterize the offer of road and public transportation (connexions, length, capacity and prices). The spatial agents refer to the different urban zones and their amenities that help to calculate zone?s attractiveness for each household or enterprise depending on its own characteristics. The attractiveness depends on several factors: the cost of housing, the environmental quality, the accessibility of urban amenities, and the supply and quality of transportation services (frequency and comfort). Global agents can decide to modify amenities in each zone (urbanism) as well as to act on road networks and public transportation offer (transport policies).
For each year, the model simulates household activity programs and calculates people displacements and good transports. It gives an overview of congestion and estimates total demand for each mode (road, public transportation and light modes). People fluxes are generated by the demand for mobility by objective and by zone on the basis of transport supply (possible distance and maximal speed). Flux intensity compared to system capacity decreases the speed of displacement. The increase in travel time is an important factor to determine modal choices (by private car, by route?s collective transports, by railway?s collective transports or by light modes). Environmental quality is partly linked to urban mobility, to displacements and to flux intensity. These factors affect the quality of the air (noises and atmospheric pollution due to transport activities) and the land use (housing, transportation networks and infrastructures, employment areas, natural spots ?) through the location choices of households and enterprises and through global agent decisions.
To test urban policies with the criteria of sustainable development, a board of indicators will be developed to present the results of city management in terms of economical, ecological and social indices. Different strategies will be implemented in the modelling shell as urban toll, public transportation development and coherent location incitements.
The project is partly funded by the French Ministry of Transport (PREDIT) and by the French Agency for Environment and Energy Management (ADEME), and the model is built in order to test the French policies on urban and transport development. It will be first applied to Rennes urban area (French Brittany). This concrete case will stage about 250.000 households (400.000 peoples) on 200 urban zones.
Association for European Transport