Abstract
Train rescheduling and rerouting issues are among the challenges faced by daily railway operations personnel in maintaining services in accordance with set timetables. These original timetables are often disrupted by unpredictable perturbation caused by a variety of incidents in the railway network such as train faults, blocked tracks and signalling faults. If the railway operator cannot maintain the services in accordance with the original schedule when these types of disruption happen, there is a strong possibility that more conflicts will arise and affect other interconnected trains in the same network. As a result, railway operators need to reschedule or reroute train actions so that the propagation of disturbances in the railway network is minimised.This thesis focuses on the development of novel solutions for Integrated Train Rescheduling and Rerouting Problems (ITRRPs) by addressing the disruption in both deterministic and stochastic scenarios. We adopted an optimisation model based on a set partitioning model taking into account safety rules and regulation principles with the aim of minimising total train delays. Metaheuristic methods were proposed and enhanced using biased randomisation to handle such ITRRPs, including the Biased Randomised Variable Neighbourhood Search algorithm (BR-VNS), Biased Randomised Hybrid VNS/TS algorithm (BR-VNS-TS), and Biased Randomised Hybrid VNS/TS with a Monte Carlo Acceptance Criterion (BR-VNS-TS-MC).
A stochastic optimisation framework was developed to deal with unexpected disruptions and to minimise total train delays. Metaheuristic methods were combined with Monte Carlo Simulation (MCS) to generate stochastic random delays consisting of Sim-Biased Randomised VNS (S-BR- VNS), Sim-Biased Randomised Hybrid VNS/TS algorithm (S-BR-VNS-TS), and the Sim-Biased Randomised Hybrid VNS/TS with a Monte Carlo Acceptance Criterion (S-BR-VNS-TS-MC) for stochastic problems.
To the best of our knowledge, this is the first time that the proposed methods have been used to solve deterministic and stochastic ITRRPs.
Experiments were conducted to evaluate the performance of the proposed algorithms using datasets from the UK South-eastern train operating company, and the State Railway of Thailand. The results showed that the proposed methods outperformed the solutions offered by existing mixed-integer linear programming (MILP) obtained by the CPLEX optimisation software package and the solution adopted by the railway operator.
Date of Award | 10 Mar 2023 |
---|---|
Original language | English |
Awarding Institution |
|
Supervisor | Banafsheh Khosravi (Supervisor) & Djamila Ouelhadj (Supervisor) |