Longitudinal model of longer-term mobility decisions: framework and first empirical tests

Recent advances in integrated land use and transport modeling have included a shift from aggregate-level modeling to disaggregate, household-level modeling. One potential advantage of this shift is that interdependencies of changes that influence household decisions can be more systematically modeled. However, existing models have not embraced this opportunity fully. Especially in the context of long-term mobility decisions (relocation/car ownership), decisions made on the basis of various dimensions are modeled as independent and cross sectional, whereas in reality they are strongly interlinked. To address these shortcomings, this paper proposes a conceptual framework that offers a more general approach to modeling the dynamics and interdependences across different time horizons of a household’s lifecycle and mobility decisions. The framework incorporates the concept of stress, defined as a discrepancy between a household’s present situation and its aspiration level, which in turn depends, among other things, on the household’s social network. Bayesian belief networks are used to represent the complex direct and indirect dependencies between life-cycle events and long- and short-term mobility decisions.