Problem:
Netherlands Railways faced a crew shortage, leading to operational delays and reduced punctuality.
Solution:
This major railway operator used an agent-based simulation model to optimize railway crew management and improve reliability. The model integrated various elements, including train timetables, crew schedules, and rolling stock plans. By simulating daily operations, comprising approximately 700 train units and over 1,000 driver duties, the company was able to evaluate different crew assignment scenarios.
Results:
- Identified critical corridors prone to delays, enabling immediate operational adjustments as a part of railway timetable planning and optimization.
- Reduced model development time to less than a week.
- Established a framework for continuous operational improvement.
Introduction: getting 9 million passengers to the right place at the right time
Netherlands Railways is the largest passenger train operator in the country, facilitating over 1 million journeys daily across a 6,900-kilometer railway network. Amsterdam Central Railway Station alone receives over 184,000 passengers every day.
Train operations are not just about trains; they involve a complex system where multiple elements must work harmoniously to ensure punctuality and efficiency. Every aspect is critical for efficient railway crew management and timetable planning.
Netherlands Railways operates one of the busiest railway networks in the world, where maintaining operational efficiency is vital to delivering reliable services for over 9 million passengers annually.
Problem: crew shortages and scheduling challenges
Following the COVID-19 pandemic, Netherlands Railways experienced a rapid growth in passenger demand from 2022 through 2023, which posed significant challenges and required railway crew management improvement.
The company needed to address several issues that could constrain its operations:
- Increasing train traffic: The railway network had to efficiently serve an increasing number of trains without expanding the existing infrastructure.
- Crew shortages: The organization experienced severe crew shortages, especially among train drivers, complicating crew management.
- Railway timetable planning: With new timetables offering more frequent services, crew schedules became increasingly tighter, leading to higher risks of delays and operational disruptions.
- Lack of powerful simulation tools: Traditional microscopic simulation tools focused only on train movements and couldn’t model crew operations and their impact on overall performance.
Solution: optimizing crew operations for better railway timetable planning
Netherlands Railways implemented a purpose-built simulation model using AnyLogic to address these challenges. The model integrated train timetables, infrastructure, rolling stock plans, and crew schedules into a cohesive simulation environment.
and overall railway crew management improvement (click to enlarge)
Read more about AnyLogic’s rail library, which allows users to efficiently model, simulate, and visualize the operations of rail yards and rail transportation of any complexity and scale.
Using agent-based modeling in AnyLogic, crew members, trains, and tasks were modeled as individual agents. Defined behaviors and interactions enabled a detailed simulation of their operations. The company designed a model to simulate the daily operations of approximately 700 train units and over 1,000 driver duties.
Netherlands Railways aimed to determine optimal buffer times between train arrivals and departures by evaluating various crew assignment scenarios for more efficient railway crew management. The simulation model provided insights into how different configurations of crew schedules affected punctuality, with specific scenarios tested, such as varying buffer times from 10 to 30 minutes.
The train operator selected AnyLogic as the primary simulation tool for its flexibility and ability to model complex systems with multiple interacting agents. Unlike traditional microscopic simulation tools, which focus only on train movements, AnyLogic allowed Netherlands Railways to integrate crew operations into a comprehensive simulation model.
The success of the AnyLogic model relied on accurate data integration. Netherlands Railways used a combination of SQL Server and Python scripts to collect and preprocess large datasets, ensuring the following smooth data import into the simulation model:
- Timetable data: Included the train schedules, specifying departure and arrival times at each station.
- Rolling stock plan: Detailed the type and capacity of train units, specifying how many units are combined to form each train.
- Crew plan: Provided individual schedules for drivers and guards, including specific tasks, break times, and home bases, while adhering to labor regulations.
The results of each simulation were exported to Power BI for visualization, making it possible to generate reports and track performance metrics across multiple runs. This integration provided actionable insights and improved communication between planning and operational teams.
Read also about another project of Netherlands Railways, focused on rail transportation network planning.
Results: achieving efficient railway crew management
The AnyLogic simulation model helped to significantly improve crew operations management and decision-making, as Netherlands Railways became the first train operator in the European Union to model crew operations comprehensively.
Netherlands Railways enhanced the reliability of crew schedules by optimizing buffer times and simulating various scenarios. They identified critical spots in crew schedules and evaluated future timetables to prepare for upcoming operational challenges.
The company’s innovative approach to leveraging simulation for railway crew management and timetable planning established a framework for continuous improvement.
Recently, Netherlands Railways has also successfully expanded the use of AnyLogic beyond crew operations to other areas, such as the placement of maintenance facilities and train cleaning optimization.
Looking ahead, the largest passenger train operator in the Netherlands plans to use AnyLogic to develop clear, quantifiable output metrics for crew plan robustness. Future work will also focus on simulating the impact of train cancellations across the network and integrating additional real-time data for even more accurate predictions.
The case study was presented by Camiel Simons from Netherlands Railways at the AnyLogic Conference 2024.
The slides are available as a PDF.
