Saturday 08 March 2025
Railway crew scheduling, a complex problem that has been puzzling transportation experts for decades. The challenge lies in optimizing the duties of train drivers to minimize cancellations and delays while ensuring they get enough rest and breaks. A new approach to tackling this issue has been developed by researchers, who have created an algorithm that can quickly and efficiently reschedule crew duties.
The traditional method of solving this problem is through a process called column generation. This involves generating all possible duties for each driver and then selecting the best combination that meets the operational constraints. However, as the number of drivers and tasks increases, so does the computational time required to solve the problem. In fact, it can take hours or even days to generate all possible duties.
Enter the tabu search approach. This method involves exploring a range of solutions by applying simple rules, such as swapping two tasks or moving an entire group of tasks to another driver. The algorithm then evaluates each solution and selects the best one that meets the operational constraints. Unlike column generation, tabu search can quickly find good solutions even for large-scale problems.
The researchers tested their algorithm on a real-world scenario provided by Mälartåg, a railway company in Sweden. They created nine instances of different sizes, ranging from 38 to 552 tasks, and simulated the absence of drivers due to various reasons such as sick leave or vacation. The results showed that the tabu search approach outperformed column generation in terms of computational time and space usage.
In addition, the algorithm was able to achieve an assignment rate of up to 88% for unassigned tasks, which is a significant improvement over traditional methods. This means that fewer trains were cancelled or delayed due to inadequate crew resources.
The researchers also explored the performance of their algorithm under different scenarios. They found that as the number of absent drivers increased, the computational time required to solve the problem did not increase significantly. This suggests that the algorithm is robust and can handle unexpected events such as a sudden surge in driver absenteeism.
While this new approach has shown promising results, it is not without its limitations. The researchers acknowledge that more work needs to be done to integrate additional constraints into the algorithm, such as the option of using taxis for deadheading or exploring more options when finding neighboring solutions.
Despite these challenges, the tabu search approach offers a significant improvement over traditional methods and has the potential to revolutionize railway crew scheduling.
Cite this article: “Efficient Railway Crew Scheduling with Tabu Search Algorithm”, The Science Archive, 2025.
Railway, Crew, Scheduling, Tabu Search, Algorithm, Column Generation, Computational Time, Space Usage, Assignment Rate, Optimization
