Nurse Scheduling Optimization in a General Clinic and an Operating Suite

A common problem in healthcare systems worldwide is nursing staff shortages combined with the uncertain nature of patient workloads. Assigning each available nurse to the right place at the right time to do the right job is a major concern among many healthcare organizations. In this dissertation we address the nurse scheduling problem by developing optimization models and efficient solution algorithms. First, we formulate a multi-objective binary integer programming model for the nurse scheduling problem where both nurse shift preferences as a proxy for job satisfaction and patient workload as a proxy for patient dissatisfaction are considered. Then, an integer programming NSP is developed that considers patient workload attributes such as patient workload type and duration as well as nurse preferences and hospital regulations. Various aspects of goals such as minimizing costs, patient dissatisfaction, and nurse idle times, and maximizing job satisfaction of nurses are considered in these models. A two-stage non-weighted goal programming solution approach is provided to find an efficient solution that addresses multiple objectives. We develop robust nurse scheduling models that consider patient workload variability as well as nurse preferences. Robust solution methodologies are developed to provide a step-by-step procedure for solving our multi-objective NSP model. The Analytic Hierarchy Process method is utilized to establish relative importance weights among objective functions in our multi-objective model. Numerical experiments are provided for our optimization models to express our models' and solution algorithms' efficiency and complexity. Finally, we develop mixed integer programming optimization models to assign nurses to different surgery cases in an operating suite. Daily nurse assignments are established based upon various attributes such as case specialties, procedural complexities and nurse skill level as well as lunch break assignments. To solve the nurse assignment model we apply three heuristic methods: solution pool feature, a modified goal programming approach and a set of swapping heuristics to develop results for the nurse assignment model. Also, a column generation scheme is developed to generate good schedules for the nurse assignment model. Actual data gathered from the University of Texas MD Anderson Cancer Center is used to show the efficiency of our solution methods.