The modeling of mode choices of intercity freight transportation with the artificial neural networks and adaptive neuro-fuzzy inference system

Mode choice modeling is probably the most important element of transportation planning. It affects the general efficiency of travel and the allocation of resources. The development of mode choice models has recently witnessed significant advances in many fields, such as passenger and freight transport. A large number of mathematical models have been used to model the traveler's choice of mode and destination and the shipper's choice of mode. shipment size and supply market, among others. Such models are not only becoming almost intractable but also data intensive, difficult to calibrate and update, and intransferable. These models cover a wide range of mathematical complexity and accuracy. This paper describes a new approach to mode choice of intercity freight transport modeling using artificial neural networks (ANNs) and adaptive neuro-fuzzy inference system (ANFIS) models. The new approach combines the learning ability of artificial neural networks and the transparent nature of fuzzy logic. The approach is found to be highly adaptive and efficient in investigating non-linear relationships among different variables. The adaptive neuro-fuzzy inference system model is tested on the freight transport market in Turkey, Germany, France and Austria by using information on the freight flows and their attributes. The ANNs and ANFIS models are more successful in the representation of the non-linear behavior of mode choice of intercity freight transport compared to the classical models. (C) 2008 Elsevier Ltd. All rights reserved.