Modeling And Workload Balancing Of A Tandem Call Center System Based On Markov Chain Theory

The concept of call center is derived from the 70 s' United States.After being introduced to China in the late 90 s,the call center industry has shown strong momentum of development in past ten years,which was widely used and drew attention in more and more industries,e.g.,finance,IT,logistics,e-commerce,government departments,medical.In the field of call center system operation management,many scholars did a lot of research on how to improve the running efficiency of call centers,among them,some scholars applied Markov queuing modeling method to discuss the performance optimization problem in tandem structure call center systems.In this paper,we firstly modeled the special structural call center system by a tandem queue network.We aggregated the process time of different types of customers and then the Matrix Geometric method was applied to achieve the stability probabilities and performance indicators such as the customers' average queue length,staff utilizations.Then,we conducted sensitivity analysis to investigate the impacts main system factors such as arrival rate,service rate and recall rate,buffer capacity may have on system performance.Thereafter,through numerical experiments,we validated that the aggregate approximation method is quite acceptable for evaluate the utilization indicators but the accuracy of the evaluation of queue length depended mainly on the variability of the process times.Using the performance evaluation method,we studied the optimization problem that considered the tradeoff of performance improvement and cost of training fee.Finally,an optimization algorithm was developed and verified to achieve the optimal workload reallocation solution.