Is it possible to maximise the amount of jobs per tour?

[williamrijksen]

Is it possible to maximise the amount of jobs per tour?

[reinterpretcat]

What would be the difference from the objective which maximizes total amount of assigned jobs and then minimizes amount of tours?

[williamrijksen]

I think I need a property called tripSize, which limits the amount of deliveries between the reloads. Does the library already supply this functionality?

[reinterpretcat]

tourSize counts all activities except departure/arrival, so you can't use it to limit trips between reload. There is no built-in tripSize constraint at the moment.

[williamrijksen]

tourSize counts all activities except departure/arrival, so you can't use it to limit trips between reload. There is no built-in tripSize constraint at the moment.

Thanks for your reply. How difficult will it be implementing it? Eventually I can create a PR myself to implement it

[reinterpretcat]

It is relatively simple to do with trivial approach, but simple implementation will affect performance as you would need to count how many jobs are served between different reloads. This approach brings an additional O(N) complexity for each job insertion evaluation. Internally, there are some caching approach used and some cached values can be reused (e.g. RELOAD_INTERVALS_KEY), but this would require a bit of understanding of internal details which are not well described at the moment (only as doc comments).

You can check tourSize as a reference for simple implementation: https://github.com/reinterpretcat/vrp/blob/master/vrp-core/src/construction/constraints/tour_size.rs

[williamrijksen]

Okay nice, I'll take a look this weekend. Of course it brings indeed an additional O(N) complexity, but I think it's not a big deal for my use case because it has not that much data.

[williamrijksen]

Found a solution which fits my case: I added a second item in the capacity array with the max amount of deliveries. In the demand I added a second item with 1. So on every delivery the load decreases until the max amount of deliveries is reached.