Bus operations scheduling subject to resource constraints using evolutionary optimization

Konstantinos Gkiotsalitis*, Rahul Kumar

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

7 Citations (Scopus)
8 Downloads (Pure)

Abstract

In public transport operations, vehicles tend to bunch together due to the instability of passenger demand and traffic conditions. Fluctuation of the expected waiting times of passengers at bus stops due to bus bunching is perceived as service unreliability and degrades the overall quality of service. For assessing the performance of high-frequency bus services, transportation authorities monitor the daily operations via Transit Management Systems (TMS) that collect vehicle positioning information in near real-time. This work explores the potential of using Automated Vehicle Location (AVL) data from the running vehicles for generating bus schedules that improve the service reliability and conform to various regulatory constraints. The computer-aided generation of optimal bus schedules is a tedious task due to the nonlinear and multi-variable nature of the bus scheduling problem. For this reason, this work develops a two-level approach where (i) the regulatory constraints are satisfied and (ii) the waiting times of passengers are optimized with the introduction of an evolutionary algorithm. This work also discusses the experimental results from the implementation of such an approach in a bi-directional bus line operated by a major bus operator in northern Europe.

Original languageEnglish
Article number9
JournalInformatics
Volume5
Issue number1
DOIs
Publication statusPublished - 6 Feb 2018
Externally publishedYes

Keywords

  • Bus bunching
  • Constrained optimization
  • Evolutionary optimization
  • Service reliability

Fingerprint

Dive into the research topics of 'Bus operations scheduling subject to resource constraints using evolutionary optimization'. Together they form a unique fingerprint.

Cite this