Predicting Call Center Performance with Machine Learning
In this paper we present a simulation-based machine learning framework to evaluate the performance of call centers having heterogeneous sets of servers and multiple types of demand. We first develop a simulation model for a call center with multi-skill agents and multi-class customers to sample quality of service (QoS) outcomes as measured by service level (SL). We then train a machine learning algorithm on a small number of simulation samples to quickly produce a look-up table of QoS for all candidate schedules. The machine learning algorithm is agnostic to the simulation and only uses information from the staff schedules. This allows our method to generalize across different real-life conditions and scenarios. Through two numerical examples using real-life call center scenarios we show that our method works surprisingly well, with out-of-sample fit (R-squared) of over 0.95 when comparing the machine learning prediction of SL to that of the ground truth from the simulation.
- 7.Breiman L, Friedman J, Stone CJ, Olshen RA. Classification and regression trees. CRC press; 1984.Google Scholar
- 8.Chen T, Guestrin C. Xgboost: a scalable tree boosting system. In: Proceedings of the 22nd acm sigkdd, international conference on knowledge discovery and data mining. ACM;2016. p. 785–94.Google Scholar
- 9.Ke G, Meng Q, Finley T, Wang T, Chen W, Ma W, Ye Q, Liu TY. Lightgbm: a highly efficient gradient boosting decision tree. Adv Neural Inf Proc Syst. 2017:3149–3157.Google Scholar