TU Darmstadt / ULB / TUbiblio

Computing Joint Delay Distributions for Trains in Railway Networks

Schwierczek, Kai (2016)
Computing Joint Delay Distributions for Trains in Railway Networks.
Technische Universität
Bachelorarbeit, Erstveröffentlichung

Kurzbeschreibung (Abstract)

Stochastic models for delay propagation in railway networks lead to promising results in the prediction of train delays. To respect waiting time policies, while preserving efficiency, these solutions assume that the departure and arrival times of trains are stochastically independent. We show, that the inherent error can become a significant problem. To tackle this problem, we first present formulas to calculate joint delay distributions for dependent trains in a basic structure. Using these distributions we can then calculate exact distributions for connecting trains. We then present a computational study comparing our calculation to a calculation, which uses the independence assumption, on a real world timetable of the German railway network (Deutsche Bahn AG). Our results show that, in the real world timetable, the error is negligible, but we still discuss how different structures could influence the result.

Typ des Eintrags: Bachelorarbeit
Erschienen: 2016
Autor(en): Schwierczek, Kai
Art des Eintrags: Erstveröffentlichung
Titel: Computing Joint Delay Distributions for Trains in Railway Networks
Sprache: Englisch
Referenten: Weihe, Prof. Karsten
Publikationsjahr: 2016
Ort: Darmstadt
Datum der mündlichen Prüfung: 12 Juni 2013
URL / URN: http://tuprints.ulb.tu-darmstadt.de/5204
Kurzbeschreibung (Abstract):

Stochastic models for delay propagation in railway networks lead to promising results in the prediction of train delays. To respect waiting time policies, while preserving efficiency, these solutions assume that the departure and arrival times of trains are stochastically independent. We show, that the inherent error can become a significant problem. To tackle this problem, we first present formulas to calculate joint delay distributions for dependent trains in a basic structure. Using these distributions we can then calculate exact distributions for connecting trains. We then present a computational study comparing our calculation to a calculation, which uses the independence assumption, on a real world timetable of the German railway network (Deutsche Bahn AG). Our results show that, in the real world timetable, the error is negligible, but we still discuss how different structures could influence the result.

URN: urn:nbn:de:tuda-tuprints-52046
Fachbereich(e)/-gebiet(e): 20 Fachbereich Informatik
20 Fachbereich Informatik > Algorithmik
Hinterlegungsdatum: 24 Jan 2016 20:55
Letzte Änderung: 24 Jan 2016 20:55
PPN:
Referenten: Weihe, Prof. Karsten
Datum der mündlichen Prüfung / Verteidigung / mdl. Prüfung: 12 Juni 2013
Export:
Suche nach Titel in: TUfind oder in Google
Frage zum Eintrag Frage zum Eintrag

Optionen (nur für Redakteure)
Redaktionelle Details anzeigen Redaktionelle Details anzeigen