The Impact of Delay Announcements in Many-Server Queues
with Abandonment
Mor Armony
Stern School of Business
New York University
Wednesday, March 15, 2006
4:30 - 5:45 PM
Terman Engineering Center, Room 453
Abstract:
In this work we develop methods to study the impact upon aggregate
system performance of state-dependent delay announcements to arriving
customers in a many-server queue with customer abandonment. We assume
that the queue is not visible to waiting customers, as in most
customer contact centers, when contact is made by telephone, email or
instant messaging. As a function of the announced delay customers may
balk or have new abandonment behavior. To perform a rough-cut
analysis, prior to a more detailed study, we use a fluid model, which
provides an approximate and highly simplified description of large
systems in an overloaded regime. In the fluid model, all customers are
faced with the same delay and consequently can be given the same delay
announcement. At the same time, the time-to-abandon distribution plays
a critical role.
We show that the resulting approximate description of aggregate
performance is effective by comparing to (1) a numerical algorithm
approximating the steady-state performance of an M/GI/s+GI queueing
model with a constant delay announcement and (2) simulations with
state-dependent announcements. Specifically, customers who cannot
enter service immediately are told the delay of the last customer to
enter service. Within the fluid-model framework, we find conditions
under which there exists a unique equilibrium delay, where the actual
delay coincides with the announced delay, and for a natural iteration
to converge to that equilibrium delay. We further consider the effect
of providing biased delay information, and show how the fluid model
can be applied to do further studies.
This is joint work with Nahum Shimkin and Ward Whitt.
Operations Research Colloquia: http://or.stanford.edu/oras_seminars.html