11.08.2013 Views

CALL CENTERS (CENTRES) - Faculty of Industrial Engineering and ...

CALL CENTERS (CENTRES) - Faculty of Industrial Engineering and ...

CALL CENTERS (CENTRES) - Faculty of Industrial Engineering and ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Keywords: M(n)/M(n)/s + GI system, Limited waiting times, Finite buffer, Virtual waiting<br />

times, Blocking-, Ab<strong>and</strong>on probability, Impatience rates, Asymptotic results, M(n)/M(n)/s<br />

system with waiting place dependent impatient rates, Markovian approximation<br />

87. Brusco, Michael J. <strong>and</strong> Larry W. Jacobs. Optimal models for meal-break <strong>and</strong> start-time flexibility<br />

in continuous tour scheduling, Management Science, 46 (12), 2000, 1630–1641.<br />

Abstract. A compact integer-programming model is presented for large-scale continuous tour<br />

scheduling problems that incorporate meal-break window, start-time b<strong>and</strong>, <strong>and</strong> start-time interval<br />

policies. For practical scheduling environments, generalized set-covering formulations<br />

(GSCF) <strong>of</strong> such problems <strong>of</strong>ten contain hundreds <strong>of</strong> millions <strong>of</strong> integer decision variables, usually<br />

precluding identification <strong>of</strong> optimal solutions. As an alternative, an implicit integer-programming<br />

model has been prevented that frequently has fewer than 1,500 variables <strong>and</strong> can be formulated<br />

<strong>and</strong> solved using PC-based hardware <strong>and</strong> s<strong>of</strong>tware platforms. An empirical study using laborrequirement<br />

distributions for customer service representatives at a Motorola Inc. call center was<br />

used to demonstrate the importance <strong>of</strong> having a model that can evaluate trade<strong>of</strong>fs among the<br />

various scheduling policies.<br />

Keywords: Continuous tour scheduling, Integer programming, Meal-break window, Start-time<br />

b<strong>and</strong>, Generalized set-covering, Customer service, Workforce scheduling, Production control<br />

88. Easton, Fred F. Labor requirements for multi-server multi-class finite queues. Working paper,<br />

Robert H. Brethen Operations Management Institute, School <strong>of</strong> Management, Syracuse University,<br />

Syracuse, NY, USA, June 2000.<br />

Abstract. Most <strong>of</strong> North American’s 70,000 call centers use Erlang queueing models to determine<br />

appropriate staffing levels for each period their systems operate. These st<strong>and</strong>ard queueing<br />

models (M/M/C or M/M/C/N) assume a one-stage single-queue service, FCFS priority, multiple<br />

identical servers, <strong>and</strong> exponential inter-arrival times <strong>and</strong> service times. If a call center<br />

provides two or more types <strong>of</strong> service, each with distinct mean service times, its workforce<br />

management system typically averages the interarrival <strong>and</strong> service times over all calls. These<br />

parameters are then automatically passed to a M/M/C/∞ or a M/M/C/N model to determine<br />

the minimum staff needed to assure a prescribed level <strong>of</strong> service.<br />

In this research we consider a multi-server finite Poisson queueing system that provides two<br />

types <strong>of</strong> service, classified H2/H2/C/N. We decompose the two arrival streams <strong>and</strong> show that<br />

the system is reversible, allowing efficient computation <strong>of</strong> its two-dimensional state probabilities.<br />

The model enables us to better underst<strong>and</strong> the magnitude <strong>of</strong> the staffing errors that result when<br />

st<strong>and</strong>ard Erlang models are applied to multi-class queueing systems. For example, we find that<br />

when an equal number <strong>of</strong> callers arrive for each <strong>of</strong> two classes <strong>of</strong> service, but the mean service<br />

rate for one class is nine times faster than the other, M/M/C/N can recommend a staffing level<br />

that is less than one-half the minimum number <strong>of</strong> employees actually needed to provide the<br />

desired level <strong>of</strong> service.<br />

89. Jongbloed, Geurt <strong>and</strong> Ger Koole. Managing uncertainty in call centers using Poisson mixtures.<br />

Working paper, Vrije Universiteit, Division <strong>of</strong> Mathematics <strong>and</strong> Computer Science, Amsterdam,<br />

34

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!