Abstract : The objective of the thesis is to propose a modeling of the residential behaviors of the households: the choice of change of residence (or removal), and the choice of the new location. This type of models aims at envisaging the probability that a household changes residence and its choice of new location, according to a certain number of explanatory variables. In our case, modeling is founded on the discrete choice theory (random utility approach). Research is based on the Lyons case, the model being built on the urban area of Lyon, on data of the year 1999.
In terms of results, side of residential mobility, it should be retained that the variables which explain the variation of the degree of mobility (of the rate of removal) of the households are the age of head, the number of children and the statute of occupation of housing. With regard to the households location, the preferences of the households for a zone are related on the characteristics of those and particularly to accessibility on various opportunities and services, which confirms that among the factors which influence the location behavior of the households one finds accessibilities, and thus, the assumption of the existence of a relation between transport and land use. It was also found that the characteristics of the households like age of head, income or household size have an influence on their location choices.
The study and the results of the models suggest that improvements and new directions of research are possible. Thus, on the side of residential mobility, there could be other variables which justify the housing change, like characteristics of the residence, residential environment or other characteristics of the households not observed. It would be also interesting to estimate the location model at a finer geographical level (îlot). Other alternatives of modeling of the residential decisions of the households would be to model a hierarchical structure of the choices, by a hierarchical or nested logit, or to model the life cycle, whit its components family, residential and professional, which are in interdependence (in condition of availability of necessary data).