Influence of Parking on Train Station Choice under Uncertainty for park-and-ride Users

Abstract

This paper presents a station choice model for park and ride (PnR) users based on uncertain parking attributes, such as parking search time (PST). In order to take into account uncertainty in PnR users’ choice process, a mixed logit model was developed within the framework of the discrete choice theory, the utility function in the model was established using a mean-variance approach under the cumulative prospect theory framework proposed Tversky and Kahneman [1]. A stated preference survey was designed for studying PnR users’ preference of stations, which was influenced by parking conditions at a train station. The experimental design was optimized using the D-optimality criterion. Our results show that the number that parking bays left in PnR facilities at given access time, parking cost including parking fees and fines and the variation of PST are important factors affecting PnR users’ station choice. PnR users were risk averse toward the variation of PST, which means that a PnR user is willing to choose a station with less uncertain or variation of PST.