EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

3616. Last-Minute Crew Rescheduling: Model and Heuristic Approach

Invited abstract in session WC-56: Railway Applications, stream Transportation.

Wednesday, 12:30-14:00
Room: S04 (building: 101)

Authors (first author is the speaker)

1. Liyun Yu
Department of Science and Technology, Linköping University
2. Carl Henrik Häll
Department of Science and Technology, Linköping University
3. Anders Peterson
Dept of Science and Technology, Linköping University
4. Christiane Schmidt
Department of Science and Technology, Linköping University

Abstract

We consider crew scheduling on the day before operation: because of a lack of train drivers (e.g., due to sick leave) crew members need to be rescheduled. We are given a feasible crew schedule for a subset of the drivers, a set of unassigned tasks, and a set of standby drivers with fixed start and end time of their working hours. We aim to generate a new crew schedule with as few unassigned tasks as possible --- which will in turn decrease the number of trains to be cancelled. We build a Mixed Integer Linear Programming (MILP), which we solve using the commercial solver Gurobi 11.0, and we develop a heuristic approach based on Tabu Search. We consider the same restrictions for the MILP model and our heuristic, this includes constraints on the total working hours, on break duration, maximum working time without a break etc. The data of a one-day schedule is provided by Mälartåg with 153 drivers in total, in which there are 16 standby drivers with fixed depot, start and end working time. For an instance with 11 drivers being on leave, we can assign all but 5 tasks (from a total of 37, which were unassigned after the leave), and we obtain a result in 23 seconds. We compare the results, computational time and space between our model and approach to show that our Tabu-Search-based approach can achieve a ‘good-enough’ result with less time and space needed.

Keywords

Status: accepted

Back to the list of papers