Some Queuing Terminology

To describe a queuing system, an input process and

an output process must be specified.
Examples of input and output processes are:

Situation Input Process Output Process

Bank Customers arrive at Tellers serve the
bank customers
Pizza parlor Request for pizza Pizza parlor send out
delivery are received truck to deliver pizzas

1
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 The Input or Arrival Process
Arrivals are called customers.
We assume that no more than one arrival can occur
at a given instant.

The Output or Service Process

To describe by the service time distribution
Servers are in parallel if all server provide the same
type of service and a customer need only pass
through one server to complete service.

2
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 8.2 Modeling Arrival and Service
Processes
ti = time at which the ith customer arrives.
Interarrival times between customer i-1 and i are
independent, continuous random variables Ai
Identical distribution of the Ai assumption of
stationary interarrival times.

3
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 We define to be the arrival rate, which will have
units of arrivals per hour.
The most common choice for A is the exponential
distribution: density
An exponential distribution with parameter has a
1 1
density a(t) = e (mean:
-t E ( A)

; variance:var A

) 2

Lemma 1: If A has an exponential distribution, then

for all nonnegative values of t and h,
P( A t h | A t ) P( A h )
no-memory property: the time until the next arrival does
not depend on how long it has been since the last arrival
4
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 Relations between Poisson
Distribution and Exponential
Distribution
If interarrival times are exponential with parameter
the number of arrivals to occur in an interval of length
t follows a Poisson distribution with parameter t.

A discrete random variable N has a Poisson

distribution with parameter t if, for n=0,1,2,,
e t ( t ) n
P( N n ) (n 0,1, 2,...)
n!

5
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 Modeling the Service Process
Assumption: service times of different customers are
independent, identically distributed random variables Si,
having a density function s(t) with mean 1/ (hours).
is the service rate (maximum number of customers to
be processed by a single server per hour)

6
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 Birth-Death Processes
Method to analyze basic queuing systems.
We define the number of people present in any queuing system
at time t to be the state of the queuing systems at time t.
We call j the steady state (equilibrium) probability that the
system is in state j at an arbitrary point in time
A birth-death process is a continuous-time stochastic process
for which the systems state at any time is a nonnegative
integer.
Most queuing systems with exponential interarrival times and
exponential service times may be modeled as birth-death
processes.

7
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 Laws of Motion for Birth-Death
Law 1
With probability jt+o(t), a birth occurs between time
t and time t+t. A birth increases the system state by 1,
to j+1. The variable j is called the birth rate in state j. In
most queuing systems, a birth is simply an arrival.
Law 2
With probability jt+o(t), a death occurs between time
t and time t + t. A death decreases the system state by 1,
to j-1. The variable j is the death rate in state j. In most
queuing systems, a death is a service completion. Note
that 0 = 0 must hold, or a negative state could occur.
Law 3
Births and deaths are independent of each other.
8
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 Derivation of Steady-State
Probabilities for Birth-Death Processes
Key point: In the long run, the rate out of any state should
equal the rate into that state
0
Same applies
j+1 j j-1 1
to any subset
.......... ..........
j+1 j 1 0 of states!
(convenience:
j+1 j 2 1
recursive
This provides a set of balance equations: procedure)
j 1 j 1 j 1 j 1 j ( j j ) ( j 1,2,...)
11 00
Normalization j = 1 linear set of equations that can
easily be solved to find the steady state probabilities j
9
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 M/M/1 queue: Poisson arrivals,
exponential service times, single server

We define . We call the traffic intensity of

the queuing system.
We now assume that 0 < 1, and find
0 1 (0 1)
j j (1 ) (0 1)

If 1, however, the infinite sum blows up.

Thus, if 1, no steady-state distribution exists.

10
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 Expected number of customers in the
system L (<1)
j j

We find: L j j j j (1 )
j 0 j 0
j
(1 ) j j
j 0

and

L (1 )
(1 ) 2 1

11
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.
 The Queuing Formula L=W
Define W = expected time a customer spends in the
queuing system (time in line plus time in service)
Define Wq = expected time a customer spends waiting
Littles formula:
L = W
Lq = Wq
where = average number of arrivals entering the system per
unit time
The queuing formula L = W is very general and can
be applied to many situations that do not seem to be
queuing problems.
12
Copyright 2004 Brooks/Cole, a division of Thomson Learning, Inc.