Trip distribution model for regional railway services considering spatial effects between stations. (15th September 2018)

• Record Type: Journal Article
• Authors: Cordera, Rubén; Sañudo, Roberto; dell'Olio, Luigi; Ibeas, Ángel
• Is Part Of: Transport policy. Volume 67(2018)
• Page Start: 77
• Page End: 84
• Publication Date: 2018-09-15
• Subjects: Railway -- Trip distribution models -- Poisson regression -- Spatial filtering; Transportation and state -- Periodicals; Transportation -- Rates -- Periodicals; 388
• DOI: 10.1016/j.tranpol.2018.01.016
• Languages: English
• ISSNs: 0967-070X

Abstract

Abstract: The railways are a priority transport mode for the European Union given their safety record and environmental sustainability. Therefore it is important to have quantitative models available which allow passenger demand for rail travel to be simulated for planning purposes and to evaluate different policies. The aim of this article is to specify and estimate trip distribution models between railway stations by considering the most influential demand variables. Two types of models were estimated: Poisson regression and gravity. The input data were the ticket sales and the prices between stations on a regional line in Cantabria (Spain) which were provided by the Spanish railway infrastructure administrator (ADIF – RAM). The models have also considered the possible existence of spatial effects between train stations. The results show that the models have a good fit to the available data, especially the gravity models constrained by origins and destinations. Furthermore, the gravity models which considered the existence of spatial effects between stations had a significantly better fit and provided a more realistic journey pattern in a future scenario than the Poisson models and the gravity models that did not consider these effects. The proposed models have therefore been shown to be good support tools for decision making in the field of railway planning. Highlights: The aim is studying trip distribution between rail stations. Poisson regression and gravity trip distribution models were estimated. Some models considered the existence of spatial effects between stations. Results showed gravity models considering spatial effects had better fit. These models showed also a more realistic journey pattern in a future scenario.