1896 IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, VOL. 16, NO. 4, AUGUST 2015
Modeling and Solving Real-Time Train Rescheduling
Problems in Railway Bottleneck Sections
Lei Chen, Member, IEEE, Clive Roberts, Member, IEEE, Felix Schmid, and Edward Stewart
Abstract—There usually exists a high density of traffic through
bottleneck sections of mainline railways, where a perturbation of
one single train could result in long consequential delays across
a number of trains. In the event of disturbances, rescheduling
trains approaching the bottleneck will be necessary to increase the
throughput of the section. To model the real-time train reschedul-
ing problems around bottleneck sections, a mixed-integer
programming model is presented in this paper. An innovative
improved algorithm (DE_JRM) is developed to solve the problem.
The model and the algorithms are validated with a case study
using Monte Carlo methodology, which demonstrates that the
proposed algorithm can reduce the weighted average delay and
satisfy the requirements of real-time traffic control applications.
Index Terms—Bottleneck section, differential evolution (DE),
railway traffic management, train rescheduling.
I. INTRODUCTION
A
S passenger and freight transport volumes increase
throughout Europe, many mainline railways have been
experiencing a high density of railway traffic, resulting in ser-
vices being very susceptible to minor delays and disturbances,
particularly in the sections where different services converge
from a range of origins and diverge to a variety of destinations.
These sections can be called “bottleneck sections.” Generally,
there are two junctions located at either end. A typical example
is the core area of the Thameslink route in London, U.K.,
as shown in Fig. 1. Due to the very high density of traffic
through bottleneck sections, the train headways are relatively
small; thus, the allowance times and the buffer times added in
the original nominal timetables are insufficient to absorb train
delays caused by unexpected disruptions that occur in daily
railway operations. A relatively small perturbation to one train
may cause long knock-on delays for the following trains on the
same route and the merging trains on other routes, because of
resource conflicts created by crossing moves and the necessary
signal overlaps.
Manuscript received June 26, 2014; revised October 28, 2014; accepted
December 2, 2014. Date of publication January 8, 2015; date of current version
July 31, 2015. This work was supported in part by the European Commis-
sion Framework Program 7 project ON-TIME and in part by the National
Natural Science Foundation of China under Grant 61304185. The Associate
Editor for this paper was B. Ning.
L. Chen, C. Roberts, and E. Stewart are with the School of Electronic,
Electrical and Computer Engineering, University of Birmingham, Birmingham
B15 2TT, U.K. (e-mail: l.chen.3@bham.ac.uk).
F. Schmid is with the School of Civil Engineering, University of Birming-
ham, Birmingham B15 2TT, U.K. (e-mail: f.schmid@bham.ac.uk).
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TITS.2014.2379617
Fig. 1. Sketch map of the core area of the Thameslink route in London, U.K.
(websites of Network Rail).
To maintain a high quality of railway services in bottleneck
sections, train delays need to be well managed to reduce their
propagation. Therefore, the trains approaching bottleneck sec-
tions need to be rescheduled in an optimal manner to minimize
the overall delays following a disturbance, which is regarded
as a key requirement of the next generation of railway traffic
management systems [1]–[3].
In a bottleneck section, there are generally two junctions
located at either end, these are referred to as known “portals.”
Approaching trains from different origins converge at the bot-
tleneck sections through the two portal junctions (arriving at
one portal junction and leaving from another one). In the bot-
tleneck sections, real-time train rescheduling problems of portal
junctions need to simultaneously consider all entry/exit points
to the bottleneck. The rescheduling of approaching trains for
one portal junction needs, therefore, to consider the movements
of the trains approaching this portal junction from another
portal junction.
The authors of this paper will present a systematic method-
ology of real-time traffic management for bottleneck sections,
including retiming and resequencing of trains when approach-
ing the bottleneck sections. The mathematical model for train
rescheduling in bottleneck sections will be introduced first. The
improved algorithm, which includes a stochastic modification
operator for solving the model, will be described in details. A
1524-9050 © 2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.