| In modern service systems, in-bound call centers are a primary channel for communication with customers. Technological impacts have led to significant changes in traditional call centers. Advancements in call center technology have enabled call centers to provide efficient service to meet customers' various needs. The skill-based routing technology enables call centers to effectively and efficiently utilize multi-skilled agents. Nonetheless, the same technologies also complicate managerial tasks, such as staffing and scheduling. New methods of management are required, which make the most effective use of these technologies.; In this research, we develop such the architecture and demonstrate its application in the case study. The subject for this case is Crutchfield, a leading mail-order catalog company selling car audio equipment and home electronics. Crutchfield's primary business channel is a call center, which current uses cross-trained agents and skill-based routing. A model based on traffic data and managerial experience from Crutchfield is examined in this study. Based on managerial recommendations, we specifically address the problem of determining call center staffing schedules, including the best routing and mix of specialized and cross-trained agents.; Instead of seeking an optimal solution, we were interested in arriving at a decision boundary for the problem. A solution architecture was proposed and developed in the research to find decision boundary information by integrating discrete-event simulation technology with tree-based classification models. The main purpose of this dissertation was to explore this solution architecture based on real call center data from Crutchfield.; Based on the research and managerial concerns for Crutchfield's call center, we accomplished call center simulation models with complex routing schemes, such as probabilistic and dynamic routing, trade-off studies for all-generalist and all-specialist call center systems in terms of efficiency and quality, and development of an integrated solution architecture by integrating discrete-event simulation, efficient sampling methods, and data mining technologies, for call center staffing. All these accomplishments resulted in concrete results for application oriented research in this dissertation. |
