Discover PyVRP Enterprise, our state-of-the-art VRP solver for large-scale route optimisation, with full source access.
PyVRP is a leading vehicle routing problem solver, built for performance, ease-of-use and extensibility.
Run our solver on your own hardware and keep your data on your own network.
Customise the solver to your own needs with direct access to the source code.
Dozens of scenarios can run in parallel on the same machine. Ideal for simulations.
Winner of the 12th DIMACS Implementation Challenge and the EURO/NeurIPS 2022 vehicle routing competition.
Break planning, reloading, multiple depots, heterogeneous fleets. See more below.
Direct consulting and implementation support from our team of routing experts.
PyVRP Enterprise handles the complex constraints real-world routing needs.
| Stop | Vehicle 1 | Vehicle 2 | Vehicle 3 |
|---|---|---|---|
| Stop 1 | |||
| Stop 2 | |||
| Stop 3 |
Balance load, duration and distance across drivers so no one driver does all the work.
Hard and soft clustering so drivers are assigned to their familiar areas.
Traffic-aware routing with travel durations that change by time of day.
PyVRP is the open-source solver behind our enterprise product. Built and maintained by the same team.
I used PyVRP to solve a large-scale real-world Salesperson Routing and Scheduling problem with complex constraints such as time windows, customer priorities, non-service days, and working-hour limits. By decomposing the problem and leveraging PyVRP for time-window routing, we consistently achieved near-optimal solutions in under a minute at production scale. PyVRP stood out for its performance, flexibility, and engineering quality, making it a highly reliable choice for practical routing optimisation problems.
We considered a fleet design problem for multi-compartment delivery vehicles, where routing is deliberately subordinate to tactical decisions. We used PyVRP as a route time estimator. It solves a TSP per customer cluster to estimate tour durations, and we also use it as a full VRP solver for benchmarking. PyVRP let us easily construct and solve hundreds of instances on the fly, embedded in our fleet design optimization loop.
The integration of PyVRP into our solutions was seamless. The chosen modelling and input formats covers a wide range of variants of the vehicle routing problem. The project ensures both compatibility with academic formats and easy applicability to problems encountered in the field.
The routing problems I get paid to solve are not found in textbooks or articles. PyVRP is great because it combines the flexibility to model a wide range of routing-like problems with the speed of being one of the most powerful VRP solvers in the market.
Working with Niels and the PyVRP team has been fantastic. They've been incredibly responsive and helpful throughout the process. The quality of their support and the professionalism they've shown makes me confident in recommending them to anyone looking for routing solutions.
PyVRP is free and MIT-licensed. PyVRP Enterprise is for teams that need additional features, deployment control, and priority support.
For engineers and researchers who need a fast, high-quality VRP solver.
Open-source, MIT licence.
View on GitHubFor teams that need additional features and priority support.
Annual source licence.
Book a technical call
