Solving vehicle routing problems with OpenStreetMap and VROOM

Julien Coupey (VERSO), State of the Map 2018
2018.stateofthemap.org/2018/T...
VROOM is an open-source software written in C++ to solve vehicle routing problems (VRP) arising in logistics and more widely in any context with geographically distributed tasks. Using OpenStreetMap data to solve real-life problems, VROOM offers an out-of-the-box integration with OSRM and is extensible to work on top of any routing engine.
The solving approach aims at providing high quality solutions efficiently by using dedicated (meta-)heuristics. This allows to get solutions very fast and/or to scale to huge problem sizes. Testing against TSPLIB, the reference benchmark for the Traveling salesman problem, shows an average optimal gap of only +2.47% while solving most instances in a few (milli)seconds.
We will describe the different ways to use VROOM by going through sample use-cases, and present the recent work toward supporting more VRP variants.

[@wexwexexort]

Great talk.

[@hasanwow]

why I am always getting 0 for distance and duration ?

[@MateusViccari]

Can we really dump google maps routing API and use a self hosting solution on which we can get faster response speeds, with a large number of points, and contribute with the community by adding missing roads? I'm in

[@johnyepthomi892]

It’s already done. You can get updated maps using Open street maps (OSM) and open street routing machine (OSRM) for routes computations. VROOM solves OR problems like VRP with multiple constraints.

[@caty863]

@@johnyepthomi892 OSRM backend is so hard to set up on windows and their documentation is lacking.