Improving Nurse Scheduling Using a Random Forest Algorithm to Predict Employee Well-Being

This paper introduces a new approach to nurse scheduling that integrates employee well-being into the decision-making process. A random forest regressor is trained to estimate a well-being score for each nurse, leveraging data from previous work weeks and considering multiple factors related to past schedules. This score is incorporated into a mixed-integer linear programming model to guide the assignment of shifts, aiming to better align schedules with individual needs. Nurses with lower well-being scores are prioritized for reduced overtime and increased shift preferences, promoting a fairer distribution of workload. The proposed method generates schedules that balance operational requirements with employee health, potentially mitigating fatigue and absenteeism.