VRP with reload facility
mertbakir opened this issue · 2 comments
Hello,
I was considering to start a similar project that has essential algorithms for VRP from the literature such as Solomon's I1, Clarke-Wright etc. Then improve over those using metaheuristics. Then I found this project yesterday. I've read the published paper, and had a chance to use the library on some toy examples.
I have 3 questions if you don't mind:
- Is a way to model cvrp with reloads a.k.a intermediary facilities, example in or-tools
Basically, I have k vehicles, single depot, n customers that requests different amounts of goods. Vehicle capacities is not sufficient to handle all the requests in a single trip. So, they need to go to the reload station to refill goods and serve the remaining customers. Refill station doesn't have to be at the depot.
- Have you considered to implement the algorithms in different language C, C++ for performance reasons?
- Do you know any similar work, heuristic algorithm implementations for VRP?
Hi there !
- CVRP with reloads is currently not available. But it would be a great enhancement. We would have to introduce new optional nodes (this is already possible); the hard part would be to reset the load resource to 0 if such nodes are visited. This is probably doable with cspy, but probably some dev is required.
- Actually, cspy (on which vrpy relies) is implemented in C++. It would be great to have everything in C or C++ (or Julia ?), but I believe it is important to have a Python API, to ensure a very user friendly interface.
- On our TO DO list, we would like to merge algorithms from VeRyPy (https://github.com/yorak/VeRyPy).
Please feel free to contribute to this project. We do not have much time nor resources to maintain it, although we have received a lot of positive feedback.