Case study
Developing a train schedule design
BNSF Railway
A railroad’s operating plan consists of railroad blocking and train scheduling. A blocking plan specifies how railcars between a specific origin-destination pair of locations are routed over the train network. A train schedule specifies how these blocks move over a rail network over a set of trains. Train scheduling determines how many trains to run, the origin-destination routes of these trains and their pickup and delivery locations. It’s a very complex optimization problem to design an optimal train schedule for a railroad, while honoring various yard and track capacity constraints. Optym partnered with BNSF Railway, the second-largest freight railroad in North America, to develop a solution for this planning challenge.
It’s a very complex optimization problem to design an optimal train schedule for a railroad, while honoring various yard and track capacity constraints.
Business problem
A railroad’s operating plan consists of railroad blocking and train scheduling. A blocking plan specifies how railcars between a specific origin-destination pair of locations are routed over the train network. A train schedule specifies how these blocks move over a rail network over a set of trains. Train scheduling determines how many trains to run, the origin-destination routes of these trains and their pickup and delivery locations. It’s a very complex optimization problem to design an optimal train schedule for a railroad, while honoring various yard and track capacity constraints. Optym partnered with BNSF Railway, the second-largest freight railroad in North America, to develop a solution for this planning challenge.
Our approach
Using a divide-and-conquer approach to create a train schedule, we broke down the original decision problem into series of smaller decision problems, then solved each decision problem optimally. We combined these solutions to solve the original problem. By employing state-of-the-art optimization techniques, including mixed integer programming (MIP), network optimization and very large-scale neighborhood (VLSN) search, we created a high-quality train schedule. Applying this approach to the data BNSF provided, we demonstrated that our train schedule optimizer could generate implementable train schedules and showed the potential benefits of plans that our optimizer generated, compared to the existing plans in use at BNSF.
Key benefits
- Since its deployment, the train schedule optimizer we created, TrainMax®, has helped BNSF save millions of dollars in operating expenses, generating an excellent return on its investment.
- BNSF uses it frequently to identify opportunities to improve its train schedule design.
- BNSF also utilizes TrainMax® to perform what-if studies, such as assessing the impact of train size on train operating costs, the impact of adding a major new yard, and the impact of train frequency on service.
Results
TrainMax® is now an integral component of BNSF’s service design suite. The railroad deploys different modules of TrainMax® to identify various types of advancements in train scheduling, such as train additions, deletions, frequency changes, time changes and block-to-train assignment changes.
