A two-stage heuristic algorithm for the nurse scheduling problem with fairness objective on weekend workload under different shift designs

Nurse rostering is a complex scheduling problem which requires the assignment of various shifts to different types of registered nurses and is subject to a set of constraints encompassing patient demand coverage and working and rest regulations. In addition to these constraints, the flexibility and fairness of shift assignments have become a pressing issue that directly affects nurse satisfaction. In this article, a two-stage heuristic algorithm is proposed to fulfill the fairness and flexibility objectives on nurse shift assignment. The first stage identifies the shift type combination that meets the demanded staffing level for each time interval of the day, and makes the shift assignments also comply with the set of working constraints. The second stage generates nurse rosters that allocate the weekend shifts as evenly as possible over a given planning horizon based on a hospital's requests. A case study using data from a US hospital is presented to illustrate the applicability of the method, and comparisons of the three most commonly used shift length designs are discussed. The results show that cost savings and fairness can be achieved through appropriate shift designs.