To quantify the effects of the introduction of new vehicles and an Internet-based realtime passenger information on Superoute 66 in Ipswich, a novel fuzzy logic model of mode choice was developed. The model is intended to capture the trade off made between the perceived levels of variables describing the cost, time comfort and convenience of a transport mode. This paper describes the framework for the model and the data required. Four input fuzzy variables describe the state of the available modes. The fuzzy inference system produces an estimate of the marginal utility of each, from which the mode share is determined using a logit model. The method of Adaptive Network Fuzzy Inference Systems is used for the calibration of the model. The available methods for determining fuzzy logic membership functions are described. It is believed that the strengths of the model lie in the ability to predict mode change behaviour that occurs when minor improvements are made to a single transport mode in a small area.
Abstract