Modello di scelta per la distribuzione delle residenze per la pianificazione della mobilità urbana

Abstract

In this scientific research, in the land use and transport interaction framework, the type of link (a correlation or causality) between the built environment and the travel behavior has been investigated: the relationship between neighborhood characteristics and travel behavior is taken into account to understand if attitudes and neighborhood preferences influence the residential location choices and travel behavior. It could be useful to examine, whether neighborhood design influences travel behavior or whether travel preferences influence the choice of neighborhood. This could lead to better understand the effect of the transport policies. A lot of studies (Hansen 1959, Lowry 1964, Nuzzolo 2006, Cao, Mokhtarian and Handy 2008) have verified that a suitable transport planning leads to an urban economics development wellmatched with the land use aims (e.g. Road Pricing, Congestion Charge and Eco-pass). The literature in transport research underlines the tools to catch the energetic sustainable town; these tools can be divided into 2 different objectives: 1) minimize the total travel time on the road network (in other terms shorten home-work distance) 2) maximize the modal split on behalf of public transport. In literature it is still not fully known which are the attitudes for the residential location choice (R.L.C.) and if between these attitudes even the travel preferences are included, and how this influences the R.L.C.. To investigate if there is a casuality or an effect between the two choices (R.L.C. and Travel choice) a survey in order to calibrate a mixed multinomial logit model in the Rome metropolitan region has been conducted. In the research, after a literature review of the land use transport models, the Experimental Design Theory will be presented and the main effect of a good survey will de discussed. It will be shown how to investigate the problem and which are the main attitudes in travel behavior and residential location choice in Rome. The Fig.1 marks the sub-area of the land use and transport system and how they are in connection. The cities in Italy are changing according to the decentralization criteria even for the household economic constraint, assuming a feature as different as possible from the old compact urban centre; new roads grow up; the buildings and the blocks are divided by streets of more than two lanes per direction and this gives an increase of the total travel time on the network for the home-work transfer. It’s easy to prove that the residential and activities replacing follows the accessibility increase or decrease. The aim of the paper is to individuate the main attributes of the utility function of the R.L.C. to evaluate the different transport policies measure to build the town to serve with an optimal transport solution. In order to obtain a result it will be necessary to make a survey on the population with an optimal experimental design - efficient designs - in order to understand the behavior of the 3 millions Roman residents with the minimum sample and with the minimum error. Through the survey it is possible to explore the connections among residential accessibility, employment, income, and auto ownership and it will be possible to focus the attention on the potential discrete choice model which will be able to predict the residential choice function of the different transport policies. A first survey has been made and the first results are useful to calibrate a choice tasks survey through the Experimental Design Theory. Later through the “Biogeme” software, the mixed utility model parameters are calibrated and with the statistical test the validity of the results are estimated. A distribution of the parameters after the calibration with biogeme is shown. For the parameters calibration even the normal, log-normal and Johnson distribution for the latent class are taken into account in order to view the differences between the outputs and underline which distribution better represents the real R.L.C., even trying to build a new Model.