Previously, multi-modal interurban trips have been modelled using, among others, a multi-nomial logit and a nested logit approach. However, the ability of these models to predict travel choice behaviour was not significantly better than a simple random assignment. The poor modelling performance may be due to both uncertainty and imprecision in traveller's perception of trip attributes. Uncertainty is caused by the non-deterministic nature of the problem. For example, travellers experience different travel times on the same trip chain each day. Imprecision, however, relates to the way travellers perceive information. Furthermore, travellers rarely have exact information. On the one hand, multi-nomial logit models have the ability to deal with uncertainty. On the other hand, it cannot deal with imprecision. Conversely, fuzzy modelling techniques may provide the answer, since these are suited to model both uncertainty and impression. In this paper a fuzzy modelling technique is applied to describe multi-modal interurban trip choice behaviour. The resulting fuzzy logic model contains, among others, access, travel, waiting, transfer and egress time and is applied separately for car and non-car owners. Preliminary results show that the fuzzy logic model outperforms the conventional multi-nomial logit model.
Abstract