A LARGE SCALE STOCHASTIC TIMETABLE-BASED TRANSIT ASSIGNMENT MODEL FOR ROUTE AND SUB-MODE CHOICES

Abstract

This paper describes how to apply a new advanced traffic assignment model for joint route and sub-mode choices, for example for different types of trains versus buses. The model is based on timetables, so that it is possible to model transfer times, hidden waiting times, delays, possibilities of obtaining a seat, etc. It is also stochastic, enabling the modelling of random variation and differences in passengers' preferences towards cost and time. It differs from many other discrete choice models because it allows for variations of the coefficients in the utility functions. It is introduced as an error component model for joint route and sub-mode choices. This model was implemented in the context of the Copenhagen-Ringsted project about the major railway between Copenhagen and the rest of Denmark. It includes a network and zone database, based on a geographical information system (GIS), and rail simulation, traffic, and economic evaluation models. To estimate the model, extensive relative preference (RP) and stated preference (SP) surveys were conducted. The steps in calibrating and validating the model are described. A table presents the model's main features: model elements and corresponding phenomena, techniques, and limitations. Some conclusions and perspectives for future work are outlined. For the covering abstract see ITRD E105584.