Solving the Nurse Rostering Problem Knowledge Representation and Reasoning Seminar

Nurse rostering is a complex problem which affects the everyday activities and health care tasks in modern hospitals. The Nurse Rostering Problem is a subclass of the personnel scheduling problems, and most of its instances are NP-hard. The goal is to find a quality assignment of shift categories to available nurses. Even though solving the Nurse Rostering Problem has been a field of interest of many research areas during many years, still the task of creating a schedule for the nurses is usually done manually. Many attempts to create an efficient algorithm for producing optimal schedules for all the available nurses in a real world hospital have been made by researchers from different areas. These attempts employ different methods in solving the Nurse Rostering Problem. Some of them combine several methods in order to obtain a hybrid one. In the following, we describe a hybrid approach for solving a real world Nurse Rostering Problem. This method makes use of Constraint Programming and Variable Neighborhood Descent in an Iterated Local Search framework. Its development has been based on the health care regulations which can be found in Norwegian hospitals. This approach has shown good test results in obtaining a feasible solution in a short amount of time.