View article

The generation of working schedules for employees in hospitals and similar institutions has been an important field of study for multiple decades. Demand for each time period must be met, while at the same time conforming to various regulations, employment contracts and employee preferences.

In the problem variant posed for the Second International Nurse Rostering Competition (INRC-II), solvers need to produce schedules for multiple consecutive and interdependent periods. In this setting, also called a stepping horizon, information about future periods becomes available only after the current schedule has been fixed. This corresponds more closely to the real-word practice found in hospitals, where demand can change within short periods, compared to the monolithic problem formulations usually studied in the literature.