Must function in developing constraints
As I have often said, the nurse scheduling problem is a mass of constraints.
Editing constraints on the scheduler nurse is a GUI (mouse or edit with a keyboard), a few dozen at most.
But if we look at the contents in detail, we can break them down into smaller constraints.
It is not uncommon to find hundreds of thousands of constraints.
That is precisely the crux of the difficulty of the nursing scheduling problem.
It is necessary to be able to point out where and what is inconsistent with such detailed constraints.
Point out the inconsistencies in the constraints.
If you get a constraint wrong, you can quickly end up with a “no solution” situation.
Even in such a case, there is a vast difference in design speed between pointing out the inconsistency and not.
It is an essential function for constraint design. It is also one of our intellectual properties in Schedule Nurse.
I have not seen any paper that mentions constraint inconsistency in optimization modeling.
It would be a real shame if modeling itself takes a great deal of time to automate the modeling process.
It is an indispensable feature for this purpose.