查看文章

The route choice phenomenon is a critical step in transportation planning process. It depends upon many factors such as travel time, distance, pavement condition etc. Route choice preferences have been mainly modelled using random utility models (RUM). These models are based on the concept of maximized utility. They assume that decision makers make decisions that maximize their utility. The problem with these models is that they may not explain total utility (observed plus unobserved). Alternative formulations have been proposed to overcome this problem in the literature. These approaches model the decision maker’s perceptions using fuzzy sets and linguistic variables through concepts from approximate reasoning and fuzzy control. Another approach is to model decision makers perception using artificial neural network (ANN) and combination of RUM plus fuzzy or ANN plus fuzzy known as neuro-fuzzy. The main idea behind the application of these soft computing techniques is that, they are capable of accounting vagueness. The claim that soft computing techniques provide better result than RUM is analysed in this paper using revealed preference (RP) survey. The route choice preference is dealt in this paper using nine route attributes (Distance, Time, Speed, Delay, Pavement condition, Parked vehicle on side of road, Aesthetic, Comfort and familiarity of route). The modelling is done using three techniques, namely multinomial logit model, fuzzy rule based inference system (FIS) and adaptive neuro-fuzzy inference system (ANFIS). It has been found that all three models have almost 100% prediction accuracy at aggregate level. At …