Bee Colony Optimization Algorithm for Nurse Rostering

In this paper, we propose a novel bee colony optimization approach to the nurse rostering problem. The bee colony optimization algorithm is motivated by the foraging habits of honey bees. In iterations, artificial bees collectively improve their solutions. The developed algorithm alternates constructive and local search phases. In the constructive phase, unscheduled shifts are assigned to available nurses, while the aim of local search phase is to improve the quality of the solution. The algorithm incorporates a novel intelligent discarding of portions of large neighborhoods for which it is predicted that they will not lead to the improvement of the objective function. Performance of the algorithm was evaluated on real-world data from hospitals in Belgium. The results show that the bee colony optimization is able to efficiently find solutions that are competitive compared to the solutions produced by other algorithms reported in the literature.