Title: Comparing time-stable performance of staffing methods using real call-centre data
Authors: Dong Dai; Arka P. Ghosh; Keguo Huang
Addresses: Department of Mathematics, Iowa State University, Ames, IA 50011, USA ' Department of Statistics, Iowa State University, Ames, IA 50011, USA ' Data Science and Analytics, FF3320-D, Bayer-Crop Science, Monsanto Company, Chesterfield, MO 63017, USA
Abstract: A central question in capacity management for service systems is to decide the number of servers that changes over time to accommodate time-varying arrivals and maintain a prescribed service-quality level. Two common methods for this are: square-root-staffing formula (SRSF) and iterative-staffing algorithm (ISA). We examine the stability of these two methods on simulated data from a probabilistic model and on a synthetic data created by resampling actual arrival, service and abandonment times from the call-centre of an Israeli bank. We use the delay probability as well as other common measures for the quality of service. In the simulated case, the ISA method marginally outperforms the SRSF method in maintaining the stability around the target delay probability. But in the case of synthetic resampled data, the stability drops when the service and patience rates are large. We also give theoretical proofs for the convergence of the ISA method under appropriate conditions.
Keywords: capacity planning; staffing; call-centres; re-sampling; data analysis; queues with time varying arrivals.
International Journal of Operational Research, 2022 Vol.44 No.1, pp.1 - 33
Received: 09 Apr 2019
Accepted: 31 May 2019
Published online: 23 May 2022 *