The value of flexibility in nurse scheduling

Abstract

Nurse scheduling is a complex class of workforce scheduling problems that involves constructing work schedules, or, assigning nurses to shifts to cover demand. Most of the studies of this problem assume fixed shifts and that demand is evenly distributed over the day. These assumptions can reduce utilization of nursing resources. In this thesis, we present a less constrained formulation of the nurse scheduling problem that minimizes staffing and preference costs while allowing intraday demand variability and staggered shifts. We developed a problem generator to create random instances and identified the most significant problem characteristics by preliminary runs. Varying and fixed shift policies were compared based on optimal solutions obtained with CPLEX. We demonstrate that nurse utilization can be improved if shift start times are allowed to vary for some types of intraday demand variability. Additionally, nurse pools with varied shift lengths were shown to improve the utilization.