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.
ecent empirical studies, namely, a time-varying arrival intensity over the course <strong>of</strong> a day, <strong>and</strong><br />
nonzero correlation between the arrival counts in different time periods within the same day.<br />
For each <strong>of</strong> the new models, we characterize the joint distribution <strong>of</strong> the vector <strong>of</strong> arrival counts<br />
with particular focus on characterizing how the new models are more flexible than st<strong>and</strong>ard or<br />
previously proposed models. We report empirical results from a study on arrival data from a<br />
real-life call center, including the essential features <strong>of</strong> the arrival process, the goodness-<strong>of</strong>-fit <strong>of</strong><br />
the estimated models, <strong>and</strong> the sensitivity <strong>of</strong> various simulated performance measures <strong>of</strong> the call<br />
center to the choice <strong>of</strong> arrival process model.<br />
Keywords: Studies, Management science, Call centers, Process engineering<br />
23. Brown, L., N. Gans, A. M<strong>and</strong>elbaum, A. Sakov, H. Shen, S. Zeltyn <strong>and</strong> L. Zhao. Statistical<br />
analysis <strong>of</strong> a telephone call center: A queueing-science perspective, JASA, 100 (469), 2005, 36–<br />
50.<br />
Abstract. A call center is a service network in which agents provide telephone-based services.<br />
Customers that seek these services are delayed in tele-queues.<br />
This paper summarizes an analysis <strong>of</strong> a unique record <strong>of</strong> call center operations. The data<br />
comprise a complete operational history <strong>of</strong> a small banking call center, call by call, over a<br />
full year. Taking the perspective <strong>of</strong> queueing theory, we decompose the service process into<br />
three fundamental components: arrivals, customer ab<strong>and</strong>onment behavior <strong>and</strong> service durations.<br />
Each component involves different basic mathematical structures <strong>and</strong> requires a different style<br />
<strong>of</strong> statistical analysis. Some <strong>of</strong> the key empirical results are sketched, along with descriptions <strong>of</strong><br />
the varied techniques required.<br />
Several statistical techniques are developed for analysis <strong>of</strong> the basic components. One <strong>of</strong> these<br />
is a test that a point process is a Poisson process. Another involves estimation <strong>of</strong> the mean<br />
function in a nonparametric regression with lognormal errors. A new graphical technique is<br />
introduced for nonparametric hazard rate estimation with censored data. Models are developed<br />
<strong>and</strong> implemented for forecasting <strong>of</strong> Poisson arrival rates.<br />
We then survey how the characteristics deduced from the statistical analyses form the building<br />
blocks for theoretically interesting <strong>and</strong> practically useful mathematical models for call center<br />
operations.<br />
Keywords: Call centers, Queueing theory, Lognormal distribution, Inhomogeneous Poisson process,<br />
Censored data, Human patience, Prediction <strong>of</strong> Poisson rates, Khintchine-Pollaczek formula,<br />
Service times, Arrival rate, Ab<strong>and</strong>onment rate, Multiserver queues<br />
(Appears also in Section I.)<br />
75