A heuristic algorithm for nurse scheduling with balanced preference satisfaction

This paper tackles the nurse scheduling problem with balanced preference satisfaction which consists of generating an assignment of shifts to nurses over a given time horizon and ensuring that the satisfaction of nurses' personal preferences for shifts is as even as possible in order to ensure fairness. We propose a heuristic algorithm based on successive resolutions of the bottleneck assignment problem. The algorithm has two phases. In the first phase, the algorithm constructs an initial solution by solving successive bottleneck assignment problems. In the second phase, two improvement procedures based on reassignment steps are applied. Computational tests are carried out using instances from the standard benchmark dataset NSPLib. Our experiments indicate that the proposed method is effective and efficient, reducing discrepancies (hence improving fairness) between the individual rosters.