Nurse Scheduling Problem

This project aims to apply learning of operations research and optimizing resources to practical cases. The aim of this problem is to maximize the the fairness of the schedule , while respecting all the constraints . In regards with the results obtained after some tests ILP has found a very good solution to our problem . Better values of the penalties associated to the shifts could be defined in order to represent the reality more accurately especially by taking into account the length of the shifts. The models can also be solved by means of optimization software. As shown in this paper, the current schedules can benefit from this work. My problem is NP-hard that it means unsolvable.My objective is to do the objective values ( diff =0.1) . The ideal must be 0 but it is impossible.there is shown in the table in 5 shifts , 20 nurses 10 weeks. We see that when we increase the number of shift the objective value is decreasing. It is fact that when we add a shift the software must do more iterations but the penalty is decreasing. We see that with the same number of iterations the penalty is smaller when we add a shift.