CALL CENTERS (CENTRES) - Faculty of Industrial Engineering and ...
CALL CENTERS (CENTRES) - Faculty of Industrial Engineering and ...
CALL CENTERS (CENTRES) - Faculty of Industrial Engineering and ...
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
95. Puhalskii, A.A. <strong>and</strong> M.I. Reiman. The multiclass GI/PH/N queue in the Halfin-Whitt regime,<br />
Advances in Applied Probability, 32 (2), 2000, 564–595.<br />
Abstract. A consideration is made <strong>of</strong> a multiserver queue in the heavy-traffic regime introduced<br />
<strong>and</strong> studied by Halfin <strong>and</strong> Whitt (1981) who investigated the case <strong>of</strong> the single customer class<br />
with exponentially distributed server times. The purpose is to extend their analysis to a system<br />
with multiple customer classes, priorities <strong>and</strong> phase-type service distributions. A weak convergence<br />
limit theorem is proven showing that a properly defined <strong>and</strong> normalized queue length<br />
process converges to a particular K-dimensional diffusion process, where K is the number <strong>of</strong><br />
phases in the service time distribution. It is also shown that a properly normalized waiting time<br />
process converges to a simple functional <strong>of</strong> the limit diffusion for the queue length.<br />
Keywords: Call Centers, Multiserver queues, Priority queues, Heavy traffic, Diffusion approximation,<br />
Weak convergence<br />
96. Reiman, Martin I. Diffusion limits for multiskill call centers with many agents. Applied Probability<br />
Society at INFORMS 2000, San Antonio, Nov. 5–8, 2000.<br />
Abstract. We consider a queueing model <strong>of</strong> a call center providing service to several customer<br />
types (skills), where each server (agent) can h<strong>and</strong>le some subset <strong>of</strong> the skills. We examine this<br />
model in the Halfin-Whitt regime, which involves the number <strong>of</strong> servers growing large while the<br />
traffic intensity approaches unity.<br />
97. Ridley, A. Performance optimization <strong>of</strong> a telecommunication call center. Proceedings <strong>of</strong> the<br />
Applied Telecommunication Symposium (ATS’00). SCS, San Diego, CA, USA, 2000, 163–167.<br />
Abstract. Telecommunication call centers have become the primary channel <strong>of</strong> customer interaction<br />
service for many businesses. The level <strong>of</strong> pr<strong>of</strong>essionalism <strong>and</strong> efficiency that call center<br />
agents deliver to customers provides a significant advantage over traditional customer service<br />
practices. The growth <strong>of</strong> call centers has been substantial over the last two decades. This growth<br />
is driven by a company’s desire to lower operating costs <strong>and</strong> to increase revenues (Kim 1997).<br />
The author investigates analytical <strong>and</strong> simulation-based models for the design <strong>and</strong> management<br />
<strong>of</strong> a call center. Given three classes <strong>of</strong> traffic (voice, E-mail, <strong>and</strong> facsimile) with different<br />
target waiting-times in queue <strong>and</strong> target service levels, the goal is to optimize the call center<br />
performance. The system performance can be measured with quantities such as the expected<br />
waiting-time in queue, the expected time in system, the percentage <strong>of</strong> calls answered within a<br />
given time, <strong>and</strong> the expected waiting-time probability distribution. The system performance <strong>of</strong><br />
the call center is measured using analytical <strong>and</strong> simulation-based queuing models. For analytical<br />
models, the traffic classes will have exponential inter-arrival <strong>and</strong> service time distributions where<br />
the arrival <strong>and</strong> service rates will differ among classes. Also, each customer call will be assigned<br />
a queue priority based on its traffic class. The call agents will be able to h<strong>and</strong>le calls from any<br />
class. For the simulation-based models, the inter-arrival <strong>and</strong> service time distributions will not<br />
be exponential, the agents will have different skill-levels, <strong>and</strong> the queue length will be finite.<br />
Keywords: Performance optimization, Telecommunication call center, Simulation-based models,<br />
37