Preferred Postal Codes

[IAH]

Hi experts,

Our customer wants to assign "preferred" postal codes to technicians in order to assign them the Orders located in those preferred postal codes, if possible. As "Preferred" is a really non-accurate word my plan is to:

1. Trigger xTour with
   • Postal code of Demand as 1 of the required Equipments for that Demand
   • Add those "Preferred postal codes" as available Equipments of the Technicians
2. Delete those PostalCodes from Required Equipments
3. Call again xTour with previous result as fixed input plan

This works fine but problem is that doing it like this we are missing the priority feature (orders of max priority could not be assigned in first iteration if its postal code is not among the preferred postal codes of the technicians).

My question is: I guess my only chance is also to iterate in this way:

1. with PostalCodes + Prio9
2. without postalCodes + Prio9
3. with PostalCodes + Prio8
4. without PostalCodes +Prio8
5. ...

Could you confirm this is the best solution?

Thank you in advance!

[kleffa]

Priority planning is used when the workload is huge, too much for the available workers. I would expect unplanned orders after planning with priorities. Otherwise priorities would be irrelevant.

In general, preferences behave similar to restrictions. Restrictions make the search space smaller so I would expect a result that is worse than without restrictions. The same holds for preferences. I would expect a result that is worse than without preferences.

In this sense, I find it a little odd to have both priorities and preferences.

But okay, here is another idea:

1. Plan with priorities but without preferences.
2. Remove the unplanned orders from the plan.
3. Plan with priorities and preferences (on the smaller set of orders, ignore the plan from the previous call)
4. Plan with priorities but without preferences with the plan from 3 as input plan.

It is really just an idea. Performance will be bad. Try with your data. I cannot tell which approach is best. I would recommend not to consider priorities and preferences in the same planning.