Estimating Schedule-Based Assignment Models for High-Speed Rail (HSR) Services Using Multiple Data Sources

The availability of open-source and big data on transport services and individual travels represents an opportunity for detailed travel demand analysis and advanced models estimation. This paper presents a schedule-based High-Speed Rail (HSR) demand forecasting model estimated using traditional and mobile phone data sources. The model includes the reconstruction of the diachronic graph of railway services provided by transport operators (both HSR, intercity and regional services) using General Transit Feed Specification (GTFS) data, and the estimation of the origin–destination (OD) matrices for each hour of the day using mobile phone traffic data. Schedule-based assignment models ultimately allow for simulating travelers’ route choices and predicting passenger flows on board single trains. The model system has been implemented for the Italian HSR network and has been calibrated and validated using ticket sales data from the primary Italian HSR transport operator, Trenitalia. This application demonstrates a high reliability of predictions, with $${R}^{2}$$ R 2 values of the comparison between observed and estimated values ranging from 0.45, when considering flows on individual train journey sections, to 0.95, when considering station-to-station flows. By presenting two practical use cases of the developed model, the results of this research highlight the importance of developing schedule-based assignment models for timetable optimization of HSR services, including their interconnections of HSR services with intercity and regional ones, and for ex-ante simulation of the impacts of policies and operational adjustments such as line extensions, speed enhancements, and the introduction of new services to improve HSR accessibility at the national scale.