|M.Sc Student||Cohen Yanir|
|Subject||Staffing and Routing Policies in Call Centers with|
|Department||Department of Industrial Engineering and Management||Supervisor||Professor Yale Herer|
|Full Thesis text|
In this thesis, we consider three related problems faced by call center companies where the ability to reassign calls to another call center is possible but incurs a cost.
We take a minimum expected cost approach and consider the following types of costs: the staffing cost, the waiting cost, and the reassignment cost. The reassignment cost includes the direct cost of routing the call to the other call center and the additional cost paid to the outsourcer; and the indirect costs that are a consequence of the difference in service quality between the two call centers (e.g., language and/or cultural issues). Thus, we are interested in finding staffing levels and a call reassignment policy that minimizes the expected total cost.
The first problem deals with a company that owns two call centers in two different locations. The company is able to reassign a percentage of calls from one call center to the other, without considering the system state when making the reassignment decision. The company would like to determine the optimal reassignment policy by finding the percentage of calls that are to be reassigned and staffing levels at each call center in order to minimize the expected total cost.
The second problem deals with a company that owns two call centers in two different locations. The company is able to reassign calls from one call center to the other. The decision on whether a call is to be reassigned is made based on the system's state (e.g., the queue length). Also here, we determine the operational decisions on staffing levels at each call center and a policy for determining when to reassign calls between the centers. For this problem we draw on an analogy to the two-location inventory problem with transshipments. In addition, we use simulation to compare and evaluate the actual savings from the above methods and to compare alternative policies.
The third problem deals with a company that owns a call center. The company is able to outsource calls to an external service provider, which can handle calls immediately. In this case, the decision on whether a call should be outsourced is made based on the system's state. The company would like to determine the optimal staffing level and an
outsourcing policy in order to minimize the expected total cost. We analyze the problem using asymptotic and exact approaches by drawing on an analogy with the M/M/N/K queue.