Estimating run choice models with innovative data collection: day-to-day tickets evolution in High Speed Rail

There are more than 110,000 km of planned and operative High-Speed Rail (HSR) lines in the world. A fundamental problem is to estimate the travel demand that uses HSR services. The paper considers models simulating the run choice of users among existing alternatives travelling between a given origin-destination pair. The research contribution concerns the proposal of a method for the identification of users from the observation of day-to-day tickets evolution. Ticket evolution constitutes a big and innovative dataset for national transportation system open to competition. The method is composed by two main parts. The former deals with the building of the choice set of alternatives, analysing the series of services tickets, and ends with the identification of user's choices. The latter phase deals with the specification and calibration of a run choice model. The proposed method has been tested on the relationship Rome-Milan (Italy), through the calibration of a disaggregated run choice model belonging to the class of random utility models. The obtained results can be important because give the possibility to update the model parameters from data obtained observing the ticket evolutions.