Performance Analysis Of Call Center Queuing System With Callback

This paper research the performance of the call center queuing system with callback.Specifically,it analyzes the TSF(Telephone Service Factor)and CR(Call resolution rate)of the system under various routing rules and agent-learning effects.The analytical method is difficult to describe the queuing system which takes into account some factors such as customer re-call and routing rule,so this paper constructs a queuing system simulation model with callback,and uses discrete event simulation method to solve performance indexes.Through the comparison and analysis of performance indexes,it is revealed that the appropriate routing rule and the emphasis on agent learning effect can significantly improve the system performance.Firstly,a call center queuing system simulation model with callback was established,then,analyze the factors that may affect TSF and CR,such as call arrival rate,agent service rate,agent number,routing rule and agent resolution rate.The results showed that:(1)Arrival rate,service rate and agent number significantly affected TSF,but did not affect CR.(2)The push-pull routing rule has a more obvious effect on the simultaneous optimization of TSF and CR than the push-pull routing rule.(3)Agent resolution rate not only significantly affects TSF,but also affects CR.Then,the influence of 6 push-pull routing rule on system performance is studied under the scenario that the agent service rate and solution rate are positively correlated,negatively correlated and irrelevant,and the system utilization rate is different.The effective use boundary of routing rule for TSF and CR is determined.The research shows that among all the routing rules,CallSwap routing rule has optimization effect on the two system performance indexes well.In particular,when the system utilization rate is no more than 90%,CallSwap routing rule can significantly improve CR while basically maintaining TSF.Finally,under the assumption that CallSwap routing strategy is adopted,the learning curve model is used to show the change of agent resolution rate,and the simulation model is established to study the influence of the learning effect of agent resolution rate on system performance under different learning rates through numerical experiments.The study shows that the learning effect significantly improves CR and TSF.