Yuri
Bizzoni,
Stefania
Degaetano-Ortlieb,
Katrin
Menzel,
Elke
Teich.
Proceedings
of
the
5th
Joint
SIGHUM
Workshop
on
Computational
Linguistics
for
Cultural
Heritage,
Social
Sciences,
Humanities
and
Literature.
2021.
Probability in the Engineering and Informational Sciences,
Год журнала:
2025,
Номер
unknown, С. 1 - 26
Опубликована: Фев. 21, 2025
In
this
paper,
we
investigate
the
number
of
customers
that
overlap
or
coincide
with
a
virtual
customer
in
an
Erlang-A
queue.
Our
analysis
starts
fluid
and
diffusion
limit
differential
equations
to
obtain
mean
variance
queue
length.
We
then
develop
precise
approximations
for
waiting
times
using
limits
polygamma
function.
Building
on
this,
introduce
novel
approximation
scheme
calculate
overlapping
customers.
This
method
facilitates
assessment
transient
risks
complex
service
systems,
offering
useful
tool
providers
mitigate
significant
overlaps
during
pandemic
seasons.
International Transactions in Operational Research,
Год журнала:
2025,
Номер
unknown
Опубликована: Апрель 21, 2025
Abstract
The
double‐ended
matching
queueing
model
has
wide
applicability
across
various
real‐life
scenarios.
In
this
paper,
we
analyze
fluid
models
of
bilateral
queues
with
time‐varying
arrivals.
contains
contents
general
patience
time
distributions
in
a
task
queue
and
worker
queue,
aiming
to
approximate
the
characteristics
systems.
We
first
prove
existence
uniqueness
solution
model.
When
arrival
rates
both
tasks
workers
are
constant,
determine
steady‐state
lengths
abandonment
proportions
for
sides
system.
Furthermore,
demonstrate
that
converges
an
equilibrium
state
starting
from
any
initial
state.
For
situations
involving
rates,
satisfactory
results
numerical
experiments
validate
reliability
describing
Finally,
explore
problems
within
systems
provide
tailored
solutions
Production and Operations Management,
Год журнала:
2021,
Номер
30(11), С. 4236 - 4257
Опубликована: Июнь 29, 2021
We
investigate
the
trade‐off
between
acquisition
and
retention
efforts
when
customers
are
sensitive
to
quality
of
service
they
receive,
that
is,
whether
get
timely
access
a
company's
resources
requested.
model
problem
as
multi‐class
queueing
network
with
new
returning
customers,
time‐dependent
arrivals,
abandonment.
derive
its
fluid
approximation;
system
ordinary
linear
differential
equations
continuous,
piecewise
smooth,
right‐hand
sides.
Based
on
model,
we
propose
novel
approach
determine
optimal
stationary
staffing
levels
for
customer
queues
in
anticipation
future
time‐varying
dynamics.
Using
accessibility
proxy
investment,
demonstrate
how
apply
our
two
families
arrival
functions
motivated
by
real‐world
applications:
an
advertising
campaign
clinical
setting.
In
numerical
study,
creates
policies
maximize
throughput
while
balancing
more
effectively
(i.e.,
equitable
abandonment
from
each
class)
than
commonly
used
near‐stationary
methods
such
variants
square‐root
policies.
Our
confirms
intimately
linked;
this
has
been
found
empirical
studies
but
not
captured
operations
literature.
suggest
environments,
focusing
either
alone
is
sufficient
maintain
high
quality.
Probability in the Engineering and Informational Sciences,
Год журнала:
2023,
Номер
38(1), С. 21 - 27
Опубликована: Фев. 23, 2023
Abstract
Imagine,
you
enter
a
grocery
store
to
buy
food.
How
many
people
do
overlap
with
in
this
store?
much
time
each
person
the
In
paper,
we
answer
these
questions
by
studying
times
between
customers
infinite
server
queue.
We
compute
closed
form
steady-state
distribution
of
pair
and
number
that
an
arriving
customer
will
with.
Finally,
define
residual
process
counts
overlapping
queue
for
at
least
$\delta$
units
its
distribution.
Frontiers in Artificial Intelligence,
Год журнала:
2021,
Номер
4
Опубликована: Июнь 1, 2021
Endeavors
to
computationally
model
language
variation
and
change
are
ever
increasing.
While
analyses
of
recent
diachronic
trends
frequently
conducted,
long-term
accounting
for
sociolinguistic
less
well-studied.
Our
work
sheds
light
on
the
temporal
dynamics
use
British
18th
century
women
as
a
group
in
transition
across
two
situational
contexts.
findings
reveal
that
formal
contexts
adapt
register
conventions,
while
informal
they
act
innovators
influencing
others.
adopted
from
other
disciplines,
our
methods
inform
(historical)
novel
ways.
These
include
periodization
by
Kullback-Leibler
divergence
determine
periods
relevant
features
variation,
event
cascades
influencer
models.
SSRN Electronic Journal,
Год журнала:
2018,
Номер
unknown
Опубликована: Янв. 1, 2018
Customer
arrival
patterns
observed
in
the
real
world
typically
exhibit
strong
seasonal
effects.
It
is
therefore
natural
to
ask:
Can
a
nonhomogeneous
Poisson
process
(NHPP)
with
rate
function
that
simple
sum
of
sinusoids
provide
an
adequate
description
reality?
If
so,
how
can
sinusoidal
NHPP
be
used
improve
performance
service
systems?
We
empirically
validate
model
consistent
arrivals
data
from
two
settings
great
interest
operations:
Patient
emergency
department
and
customer
calls
bank
call
centre.
This
finding
provides
rigorous
justification
for
use
assumption
many
existing
queuing
models.
also
clarify
why
more
suitable
than
standard
when
underlying
pattern
aperiodic
(e.g.,
does
not
follow
weekly
cycle).
illustrated
using
car
dealership,
via
naturalistic
staffing
simulation
based
on
On
other
hand,
if
periodic,
we
explain
both
models
should
perform
comparably.
Even
then,
still
necessary
managers
verify
indeed
step
seldom
performed
applications.
SSRN Electronic Journal,
Год журнала:
2018,
Номер
unknown
Опубликована: Янв. 1, 2018
We
study
customer
service
chat
(CSC)
systems
with
generally
distributed
and
patience
times
by
developing
measure-valued
processes
to
model
analyze
the
system
dynamics.
first
prove
that
these
are
tight
in
many-server
asymptotic
regime
then
show
their
limits
satisfy
a
set
of
fluid
equations.
establish
invariant
states
use
obtain
(non-asymptotic)
approximations
for
various
performance
metrics
CSC
steady
state.
also
demonstrate
accuracy
using
extensive
numerical
experiments.
These
allow
us
impact
time
distributions
on
devise
effective
dynamic
routing
policies.
SSRN Electronic Journal,
Год журнала:
2024,
Номер
unknown
Опубликована: Янв. 1, 2024
This
paper
explores
agent
multitasking
over
time
in
customer
service
environments,
particularly
within
livechat
settings.
Multitasking,
a
ubiquitous
occurrence
various
professional
sectors,
is
typically
assumed
to
have
constant
effect
conditional
on
the
level.
We
go
beyond
this
conventional
approach
by
examining
temporal
dynamics
of
and
their
effects
performance.
Through
collaboration
with
third-party
contact
center,
we
analyze
interactions
across
different
retail
brands,
focusing
both
multibranding,
where
agents
manage
tasks
for
multiple
brands
simultaneously.
Our
findings
reveal
that
handling
chats
simultaneously
increases
average
chat
duration,
diminishing
marginal
as
number
concurrent
grows.
explore
implications
these
insights
proposing
"pile-on"
heuristic,
which
assigns
new
already
engaged
multitasking,
rather
than
evenly
distributing
agents.
method
can
lead
reduction
duration
increased
idle
time.
study
contributes
operations
management
literature
illustrating
dynamic
nature
challenging
traditional
perceptions
suggesting
intensity
play
crucial
roles
shaping
operational
efficiency,
show
how
work
assignment
strategies
influence
outcomes.
2018 Winter Simulation Conference (WSC),
Год журнала:
2021,
Номер
unknown
Опубликована: Дек. 12, 2021
Processor
sharing
queues
are
used
in
a
variety
of
settings
telecommunications
and
internet
applications
known
for
being
fair.
In
this
paper,
we
study
the
possibility
accurately
predicting
response
times
real
time
G/G/1
processor
queue.
To
end,
combine
stochastic
simulation
with
supervised
learning
machine
methods.
We
show
that
many
current
algorithms
can
perform
well
at
real-time.