CALL CENTERS (CENTRES) - Faculty of Industrial Engineering and ...
CALL CENTERS (CENTRES) - Faculty of Industrial Engineering and ...
CALL CENTERS (CENTRES) - Faculty of Industrial Engineering and ...
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Keywords: Operations research, Call centers, Service level agreements, Equilibrium, Mathematical<br />
models, Routing, Optimization, Real time<br />
130. Atar, R., A. M<strong>and</strong>elbaum <strong>and</strong> M.I. Reiman. Scheduling a multi-class queue with many exponential<br />
servers: Asymptotic optimality in heavy-traffic, Annals <strong>of</strong> Appl. Prob., 14 (3), 2004,<br />
1084–1134. Downloadable from: .<br />
Abstract. We consider the problem <strong>of</strong> scheduling a queueing system in which many i.i.d.<br />
servers cater to several classes <strong>of</strong> impatient customers. Service times <strong>and</strong> impatience clocks are<br />
exponential while arrival processes are renewal. Our cost is an expected cumulative discounted<br />
function, linear or nonlinear, <strong>of</strong> appropriately normalized performance measures. As a special<br />
case, the cost per unit time can be a function <strong>of</strong> the number <strong>of</strong> customers waiting to be served<br />
in each class; the number actually being served, the ab<strong>and</strong>onment rate, the delay experienced<br />
by customers, the number <strong>of</strong> idling servers, as well as certain combinations there<strong>of</strong>. We study<br />
the system in an asymptotic heavy-traffic regime where the number <strong>of</strong> servers n <strong>and</strong> the <strong>of</strong>fered<br />
load R are simultaneously scaled up <strong>and</strong> carefully balanced: n ≈ R + β √ R, for some scalar β.<br />
This yields an operation that enjoys the benefits <strong>of</strong> both heavy traffic (high server utilization)<br />
<strong>and</strong> light traffic (high service levels.)<br />
We first consider a formal weak limit, through which our queueing scheduling problem gives rise<br />
to a diffusion control problem. We show that the latter has an optimal Markov control policy,<br />
<strong>and</strong> that the corresponding Hamilton-Jacobi-Bellman (HJB) equation has a unique classical solution.<br />
The Markov control policy <strong>and</strong> the HJB equation are then used to define scheduling<br />
control policies which we prove are asymptotically optimal for our original queueing system.<br />
The analysis yields both qualitative <strong>and</strong> quantitative insights, in particular on staffing levels,<br />
the roles <strong>of</strong> non-preemption <strong>and</strong> work-conservation, <strong>and</strong> the trade<strong>of</strong>f between service quality <strong>and</strong><br />
servers’ efficiency.<br />
131. Atar, R., A. M<strong>and</strong>elbaum <strong>and</strong> M.I. Reiman. A Brownian control problem for a simple queueing<br />
system in the Halfin-Whitt regime, Systems <strong>and</strong> Control Letters, 51 (3–4), 2004, 269–275.<br />
Abstract. We consider a formal diffusion limit for a control problem <strong>of</strong> a multi-type multiserver<br />
queueing system, in the regime proposed by Halfin <strong>and</strong> Whitt, in the form <strong>of</strong> a control<br />
problem where the dynamics are driven by a Brownian motion. In one dimension, a pathwise<br />
minimum is obtained <strong>and</strong> is characterized as the solution to a SDE. The pathwise solution to a<br />
special multi-dimensional problem (corresponding to a multi-type system) follows.<br />
Keywords: Queueing networks, Stochastic control, Heavy traffic asymptotics<br />
132. Atlason, Julius, Marina A. Epelman <strong>and</strong> Shane G. Henderson. Call center staffing with simulation<br />
<strong>and</strong> cutting plane methods, Annals <strong>of</strong> Operations Research, 127 (1–4), March 2004, 333–358.<br />
Available at: http://critical.orie.cornell.edu/∼shane/pubs.html.<br />
Abstract. We present an iterative cutting plane method for minimizing staffing costs in a<br />
service system subject to satisfying acceptable service level requirements over multiple time periods.<br />
We assume that the service level cannot be easily computed, <strong>and</strong> instead, is evaluated<br />
49