Recent studies of urban travel are showing patterns of personal movement that would not be correctly predicted by conventional travel models. Examples are given of the stability of total trip numbers and of travel time budgets, of the restricted availability of private cars and of the lack of free choice between modes. Conventional and disaggregated models are examined and it is shown that accurate calibration is no guarantee of accurate predictive ability. Current trip generation models are considered to be satisfactory, and recent advances in modal split and trip distribution models are described. The fundamental question of the spatial and temporal stability of transport models, and hence the validity of these models as predictive tools, is discussed. It is concluded that the combinations of modelling with more disaggregated data, of incorporating existing improvements to modal split and trip distribution models, of allowing for restricted availability of some modes and other constraints on travellers' behaviour and of including the constraints and invariant factors noted in recent studies, will indeed lead to the development of transport models that do reflect the real world. (Author/publisher)
Samenvatting