Variable employee productivity in workforce scheduling

This paper considers the problem of developing workforce schedules using groups of employees having different productivity. We show that the existing linear representation of this problem is often inaccurate for high-contact service organizations because it ignores the stochastic nature of customer arrivals. Specifically, the existing representation commonly overestimates the number of less productive employees necessary to deliver a specified, waiting time-based customer service level. We present a new, nonlinear representation of this staffing problem that captures its nonlinear nature and demonstrate its superiority via an extensive set of labor tour scheduling problems for the two-group case.