Hilditch, Andrew (2022) Data-driven network analysis and delay study of the British Railway. PhD thesis, University of Warwick.

Preview

PDF WRAP_Theses_Hilditch_2022.pdf - Submitted Version - Requires a PDF viewer. Download (6Mb) | Preview

Request Changes to record.

Abstract

In this thesis we look at the British railway network and perform data-driven network analysis on publicly available data. We perform centrality analysis as well as basic network theory methods on different levels of the network. The analysis allows us to see that the historic hubs can be found from current data, and we also located some of the modern hubs. We then move onto an in-depth case study of a local line where we examine bottlenecks and learn that the section of line nearest to London has the most issues and discover that turn-around times at terminuses pose an obstacle to catching back up from delay. We later propose a formula to estimate outbound delay with respect to inbound delay. Two Random Forest Regression methods are used in this thesis. The first is used to guess how delayed a train will be at a station and is used to find the feature importance measures. We find that the time of day and the station being predicted are the two factors with the greatest contribution. The second regression method is used to predict travel times to stations. This method is compared to a baseline prediction and outperforms it easily. Work is also done to examine train interactions and knock-on delay. We propose that trains on the line can form \concertinas", a relationship where trains move closer and then further apart, this delays the rear train. These relationships are studied in the form of interval graphs.

Item Type:	Thesis (PhD)
Subjects:	Q Science > QA Mathematics T Technology > T Technology (General) T Technology > TF Railroad engineering and operation
Library of Congress Subject Headings (LCSH):	Railroads -- Joint use of facilities, System analysis, Network analysis (Planning), Train delays
Official Date:	November 2022
Dates:	Date Event November 2022 UNSPECIFIED
Institution:	University of Warwick
Theses Department:	Mathematics for Real-World Systems Centre for Doctoral Training
Thesis Type:	PhD
Publication Status:	Unpublished
Supervisor(s)/Advisor:	Connaughton, Colm
Sponsors:	Engineering and Physical Sciences Research Council ; Thales Engineering & Consulting
Format of File:	pdf
Extent:	108 pages : illustrations
Language:	eng

Request changes or add full text files to a record

Repository staff actions (login required)

View Item

Downloads