A BI-OBJECTIVE MATHEMATICAL MODEL FOR NURSE SCHEDULING PROBLEM CONSIDERING NURSE'S PREFERENCES BASED ON DEA APPROACH

The nurse scheduling in a hospital is a complex and time-consuming problem which considers assigning nurses to shifts for each day of a planning horizon while ensuring meeting the demand of hospital units. In developing countries, there is usually a shortage of nursing staff in health centers' therefore, the nurse scheduling problem is one of the most important issues in human resource management in clinical units. In this research, a mathematical model is developed so that constraints are classified into two types of hard and soft and the weight of soft ones is obtained using the pairwise comparison matrix. In the proposed model, two objective functions are considered to maximize nurses' preferences and minimize the deviations from soft constraints for nursing scheduling problems. The nurses' preferences represent a very important issue in nurses' satisfaction. As a novelty of this paper, three factors used to calculate the nurses' preferences based on the data envelopment analysis (DEA) method are as follows: nurses' preferential ratings, data related to the preferences of past scheduling periods, and the work experience of nurses. Hospital nurses are also divided into two groups: fixed shift work and rotational shift work. Also, a fair allocation is considered for night and weekend shifts for nurses. The proposed model was solved by an improved version of the augmented epsilon constraint method (AUGMECON2) using the data for a case study in the Intensive Care Unit (ICU) in Loghman Hakim Hospital in Tehran, Iran. Comparing the results of the solution of the proposed model with the current method shows that there is a significant improvement in preparing the nursing timetable and responding to nurses' preferences. The computational results of the mathematical model show that the nurses' mandatory overtime is reduced' therefore, the hospital costs are decreased. Also, a sensitivity analysis is presented for the deviations from soft constraints with respect to maximum working hours.