Abstract
The Nurse Rostering Problem (NRP) is a task of assigning duties fairly among available nurses at healthcare organizations for specific periods of time defined as shifts. NRP belongs to NP-hard problems for which an optimal solution is difficult to be obtained. This paper reports on the practical nurse rostering problem and issues at Hospital Universiti Kebangsaan Malaysia (HUKM). A survey (questionnaire and interviews) was conducted and used as the main tool for gathering primary data concerning the nurses' preferences and other constraints. Furthermore, policy maker views were collected and analyzed. The results identified the majority of rostering problems that the nurses are currently facing at HUKM. The results further revealed that the underlying need for developing an automated fair nurse roster is increasingly becoming critical. Based upon the results, a mathematical model for HUKM nurse rostering problem was formulated. A new objective function was also presented to cope with such a real world problem. Additionally, a new dataset for HUKM was gathered which is readily available for future researches and studies.