Pareto-optimal workforce scheduling with worker skills and preferences

This paper addresses employee scheduling in service operations, considering various skill and skill levels and the fluctuating customer demand throughout the day and week. Employee shift and day-off preferences are also considered to enhance morale. We propose a two-stage integer programming model. In the first stage, the model optimizes the number of employees required for each shift period, ensuring uniform distribution of overstaffing to improve customer service. A Pareto frontier approach is applied between the two stages, offering decision-makers a set of non-dominated solutions that balance overstaffing and understaffing. The second stage uses the selected Pareto-optimal solution to assign shifts and day-offs to employees, incorporating their skills, preferences, and fairness considerations. Our model implicitly includes shifts and breaks, reducing decision variables and computational time. Using real data from a dining restaurant chain, we validate the model's effectiveness in enhancing customer service and reducing labor costs by 12.3% compared to manual scheduling. Furthermore, productivity and employee satisfaction improve by considering individual skills and preferences.