Telephone Call Centers - Performance, Design and Control of Time Varying Queues

点击翻译按钮获取中文摘要

This research plan aims to investigate four specific areas: (1) the analysis of sojourn time distributions for call centers; (2) the modelling of contact centers; (3) server staffing for call centers with finite buffers and impatient customers; and (4) server staffing with non-Gaussian diffusion limits for some Markovian service network models. The time varying queueing models to be developed through their asymptotic analysis are guided by a desire to create an operational analysis for call centers. The results will address the fundamental problems of performance, design and control for these systems. One ultimate goal of the methods is to create new simplifying algorithms that practitioners can use. These new mathematical techniques will increase the understanding of time-dependent behavior and give tools for creating new types of call centers. New "rule of thumb" formulas and algorithms are needed for these challenges. Developing a deeper analysis for a broader class of queueing models becomes an important way to provide this insight. The majority of interactions between major businesses and their customers happens through the communication medium of call centers. This transfer of information helps businesses prosper by increasing the sales of their products and getting the necessary feedback to be more responsive to customer needs. Thus, the growth of call centers is driven by the growth of new businesses, the expansion of old ones, as well as the speed, capacity and cost of information technology. Nearly two-thirds of the cost for running call centers involves issues of staffing. Hence any methodology that optimizes the staffing of call centers has a significant impact on the efficient management of businesses. Traditionally queueing theory has been an important tool in the design and performance of service systems. However, these traditional queueing models do not begin to describe all the detailed attributes found in the real world complexity of call centers. Such systems can easily have time varying numbers agents, heterogeneous types of customers needs and multiple skill sets for agents. Each of these features points to a need to develop a richer class of tractable queueing models for the operational design, performance and control of call centers.