Relationships between Level and Weight

Q.

Would you please be so kind as to explain me what the purpose of the

"weight" in the Solve tab ?

Is there any impact changing the integer value of it ?

I understand the "soft level" principle but the weight...

I may have missed this part somewhere in the documentation, sorry if it is

the case.

Kind regards,

Ans.

For more information on the relationship between levels and weights,

Please read the following section.

https://www.nurse-scheduling-software.com/manuals/shift_manuals/chapter19/

You should see ******UB=... in the right pane in the fig.

The objective function value (UB) of a linear system is,

The following equation.

UB = ΣWeigts[i]*Errors[i].

The solver's task is then to find the shift assignment that minimizes the value of UB.

Optimization in linear systems is to minimize UB.

So, the magnitude of the weights corresponds to the priority. The larger the weight, the higher the priority.

In contrast, a level is like an index that summaries the weights.

By convention, level 7 is marked as a serious error, with red and others marked yellow, but there is no relationship between that and the weights. It is also fine if level 7 is weighted 1 and level 1 is weighted 10. It is only the weights that are passed to the solver.

However, the advantage of such a system for Scheduling Nurse is that the weights of the desired constraints can be changed immediately at hand when seeking solutions. As is well known, the nurse scheduling problem is not known what the error looks like at the time until it is solved. Depending on the state of the error at the time, the priority can be changed immediately at hand.