Nurse scheduling with multi-grades and individual preferences

The compilation of effective and satisfactory nurse schedules plays an important role in reducing the cost and increasing the nursing quality in a hospital. Hence, a nurse scheduling problem with multi-grades and individual preferences is studied. A set of potential shift patterns are first built, based on which an integer programming model is constructed. Next, considering the characteristics of nurse schedules, two hybrid two-stage nurse scheduling approaches are developed, in which a random-matching based heuristic is first devised to make the initial schedule feasible, and then, two meta-heuristics, such as simulating annealing and tabu search are tailored to minimize cost further. Experiments show that the approaches are effective and considerably better than simple heuristics in terms of solution quality.