Dear Johan, Kai,
Seems that right now carrier vehicle types can
either be defined in the carrier plans file
or be defined in a separate file.
I don't think that we want to support both execution paths. Do you have a preference?
(In terms of coding I don't think that it makes a difference. In terms of maintenance, it might be easier to have it in separate files, because one might be able to bump up the version of the plans file without having to bump up the version of the vehicleTypes file, and vice versa.)
Note that the carrier vehicles themselves will remain in the carrier plans file; I think that that would be too much of a change.
Thanks and best wishes
[Created via e-mail received from: "Nagel, Kai, Prof. Dr." <firstname.lastname@example.org>]
We here in Berlin (TU and DLR) are using option 2) separate file for reading in the vehicleTypes.
So for me it would be totally fine to remove option 1) vehicleType definition in the Carrier plans file.
Regarding the carrier vehicles: I am in favour of let them remain in the carriers file, because they are necessary for the definition of the VehicleRoutingProblem and define the fleet (infinite or finite) of the carrier.
I will prefer it in the plans file, because the separate file never made sense to me. In the plans file makes more sense if you script the population synthesis, while a separate file means there is always one more file that, to get the code running after a few months, you’re not so sure WHAT VERSION is the latest.
There are, yes, ways to overcome my frustration above… but my preference would be for “fewer files”.
But I would gladly accept if, Kai, you prefer code-maintainability over “fewer files”.