Α two-phase adaptive variable neighborhood approach for nurse rostering

Abstract

This contribution presents a two-phase variable neighborhood search algorithm for solving nurse rostering problems. In order to demonstrate the efficiency of the proposed algorithm, it is firstly applied to all nurse rostering problem instances as proposed in the First International Nurse Rostering Competition (INRC-2010). Computational results assessed on all three sets of sixty competition instances demonstrate that the proposed algorithm improves the best known results for two instances, inside the time limits of the competition, while achieving the best known bounds for forty eight other instances. The proposed algorithm was also applied to seven other nurse rostering instances reported in the respective literature and managed to achieve the best known result in six of them while improving the best known result in one instance. The proposed algorithm, as well as its differences from existing approaches are presented, described and discussed in detail.