Modelling user mode choices by an ellipsoidal fuzzy approach

Among transport-related choices one of the most important and studied is the travel mode choice. In fact, the estimation of the travel mode percentages is crucial to forecast possible configurations of the transportation system according to suitable policies. In this work, the travel mode choice is studied by using a fuzzy approach. The underlying idea is that user travel behaviour is intrinsically uncertain and ambiguous and the user choices can be predicted only to some extent. The fuzzy approach aims at identifying the most important variables affecting user’s mode choices and predicts mode choices on the basis of fuzzy similarities approach based on ellipsoidal rules and neuro-fuzzy tuning procedure. The procedure has been tested by using a large database concerning actual mode choices outperforming other conventional fuzzy systems exploiting generalized network-based fuzzy inference.