Constraint satisfaction problem based modelling and value ordering based heuristic for nurse scheduling

Abstract

Nurse Scheduling Problem (NSP) consists of assigning shift-types (morning, afternoon and night) to qualified personnel over a certain planning period. It is a difficult and time-consuming task. In this paper we present a formulation of the hospital Nurse Scheduling Problem just like Constraint Satisfaction Problem ′′CSP′′ based constraint programming in order to find a solution, which minimizes the violation of Nurses’ preferences. We would suggest a flexibility tool for helping to decide and for making negotiation easier. Our originality lies in the modelling of the problem, by defining the global constraints and in the algorithm of resolution to solve it, by proposing a new value ordering based heuristic for assignment of the decision variables taking into account the set of Nurses preferences. Our heuristic is based on the structure of the CSP and on the properties of constraints. It allows the reduction of the search space for solution and returns a solution within few second.