Chapter 17

Waiting Line Analysis

for Service Improvement
Operations Management - 5th Edition

Roberta Russell & Bernard W. Taylor, III

Beni Asllani
Copyright 2006 John Wiley & Sons, Inc. University of Tennessee at Chattanooga
 Lecture Outline

 Elements of Waiting Line Analysis
 Waiting Line Analysis and Quality
 Single Server Models
 Multiple Server Model

Copyright 2006 John Wiley & Sons, Inc. 17-2

 Waiting Line Analysis

 Operating characteristics
 average values for characteristics that describe the 
performance of a waiting line system
 Queue
 A single waiting line
 Waiting line system consists of
 Arrivals
 Servers
 Waiting line structures

Copyright 2006 John Wiley & Sons, Inc. 17-3

 Elements of a Waiting Line

 Calling population
 Source of customers
 Infinite ­ large enough that one more customer can 
always arrive to be served
 Finite ­ countable number of potential customers
 Arrival rate (λ)
 Frequency of customer arrivals at waiting line 
system 
 Typically follows Poisson distribution

Copyright 2006 John Wiley & Sons, Inc. 17-4

 Elements of a Waiting Line (cont.)

 Service time
 Often follows negative exponential 
distribution
 Average service rate = μ
 Arrival rate (λ) must be less than service 
rate (μ) or system never clears out

Copyright 2006 John Wiley & Sons, Inc. 17-5

 Elements of a Waiting Line (cont.)

 Queue discipline
 Order in which customers are served
 First come, first served is most common
 Length can be infinite or finite
 Infinite is most common
 Finite is limited by some physical

Copyright 2006 John Wiley & Sons, Inc. 17-6

 Basic Waiting Line Structures

 Channels are the number of parallel servers
 Single channel
 Multiple channels
 Phases denote number of sequential servers the 
customer must go through
 Single phase
 Multiple phases
 Steady state
 A constant, average value for performance characteristics that 
system will reach after a long time

Copyright 2006 John Wiley & Sons, Inc. 17-7

 Operating Characteristics
NOTATION OPERATING
CHARACTERISTIC
L Average number of customers in the system (waiting and
being served)
Lq Average number of customers in the waiting line
W Average time a customer spends in the system (waiting and
being served)
Wq Average time a customer spends waiting in line
P0 Probability of no (zero) customers in the system
Pn Probability of n customers in the system
ρ Utilization rate; the proportion of time the system is in use
Table 16.1

Copyright 2006 John Wiley & Sons, Inc. 17-8

 Cost Relationship in Waiting
Line Analysis
Total cost

Service cost
Expected costs

Waiting Costs

Level of service
Copyright 2006 John Wiley & Sons, Inc. 17-9
 Waiting Line Costs and Quality
Service
 Traditional view is that the level of 
service should coincide with minimum 
point on total cost curve
 TQM approach is that absolute quality 
service will be the most cost­effective in 
the long run

Copyright 2006 John Wiley & Sons, Inc. 17-10

 Single-server Models

 All assume Poisson arrival rate
 Variations
 Exponential service times
 General (or unknown) distribution of service times
 Constant service times
 Exponential service times with finite queue length
 Exponential service times with finite calling 
population

Copyright 2006 John Wiley & Sons, Inc. 17-11

 Basic Single-Server
Model: Assumptions
 Poisson arrival rate
 Exponential service times
 First-come, first-served queue discipline
 Infinite queue length
 Infinite calling population
 λ = mean arrival rate
 µ = mean service rate

Copyright 2006 John Wiley & Sons, Inc. 17-12

 Formulas for Single-
Server Model
Probability that no customers
are in the system (either in the P0 = 1 - λ
queue or being served) µ
n
λ
Pn = • P0
Probability of exactly n µ
customers in the system n
λ
= 1- λ
µ µ
Average number of λ
L =
customers in the system µ - λ

Average number of λ2
Lq =
customers in the waiting line µ (µ -
λ)
Copyright 2006 John Wiley & Sons, Inc. 17-13
 Formulas for Single-
Server Model (cont.)
Average time a customer 1 L
W = =
spends in the queuing system µ - λ
λ
Average time a customer
λ
spends waiting in line to Wq =
be served µ (µ -
λ)
Probability that the server
λ
is busy and the customer ρ =
has to wait µ

Probability that the server I = 1-ρ

is idle and a customer can λ
be served = 1- = P0
µ
Copyright 2006 John Wiley & Sons, Inc. 17-14
 A Single-Server Model

Given λ = 24 per hour, µ = 30 customers per hour

Probability of no
P0 = 1 - λ
24
customers in the = 1- =
system µ 30
0.20

Average number
λ 24
of customers in L = = =4
µ -λ 30 - 24
the system

Average number (24)2

λ2
of customers Lq = = = 3.2
µ (µ - λ ) 30(30 - 24)
waiting in line

Copyright 2006 John Wiley & Sons, Inc. 17-15

 A Single-Server Model

Average time in the 1 1

W = = = 0.167 hour
system per customer µ - 30 - 24
λ
Average time waiting λ 24
Wq = = = 0.133
in line per customer µ (µ - λ ) 30(30 - 24)

Probability that the

λ 24
server will be busy and ρ = = = 0.80
the customer must wait µ 30

Probability the
I = 1 - ρ = 1 - 0.80 = 0.20
server will be idle

Copyright 2006 John Wiley & Sons, Inc. 17-16

 Service Improvement Analysis

 Possible Alternatives
 Another employee to pack up purchases

 service rate will increase from 30 customers to 40 
customers per hour
 waiting time will reduce to only 2.25 minutes

 Another checkout counter
 arrival rate at each register will decrease from 24 to 12 per 
hour
 customer waiting time will be 1.33 minutes

 Determining whether these improvements are worth 
the cost to achieve them is the crux of waiting line 
analysis
Copyright 2006 John Wiley & Sons, Inc. 17-17
 Constant Service Times

 Constant service times 
occur with machinery 
and automated 
equipment
 Constant service times 
are a special case of the 
single­server model with 
undefined service times

Copyright 2006 John Wiley & Sons, Inc. 17-18

 Operating Characteristics for
Constant Service Times

Probability that no customers P0 = 1 - λ

are in system µ

Average number of λ2
Lq =
customers in queue 2 µ (µ - λ )

Average number of λ
L = Lq +
customers in system µ

Copyright 2006 John Wiley & Sons, Inc. 17-19

 Operating Characteristics for
Constant Service Times (cont.)

Average time customer Lq

spends in queue Wq =
λ

Average time customer 1

spends in the system W = Wq +
µ

Probability that the λ

ρ =
server is busy µ

Copyright 2006 John Wiley & Sons, Inc. 17-20

 Constant Service Times:
Example
Automated car wash with service time = 4.5 min
Cars arrive at rate λ = 10/hour (Poisson)
µ = 60/4.5 = 13.3/hour

λ2 (10)2
Lq = = = 1.14 cars waiting
2µ (µ - λ ) 2(13.3)(13.3 - 10)

Lq
Wq = = 1.14/10 = .114 hour or 6.84 minutes
λ

Copyright 2006 John Wiley & Sons, Inc. 17-21

 Finite Queue Length
 A physical limit exists on length of queue
 M = maximum number in queue
 Service rate does not have to exceed arrival rate (µ > λ)
to obtain steady-state conditions

Probability that no 1 - λ /µ
P0 =
customers are in system 1 - (λ /µ )M + 1
n
Probability of exactly n λ
customers in system Pn = (P0) for n ≤ M
µ

Average number of λ /µ (M + 1)(λ /µ )M + 1

L= -
customers in system 1 - λ /µ 1 - (λ /µ )M + 1

Copyright 2006 John Wiley & Sons, Inc. 17-22

 Finite Queue Length (cont.)

Let PM = probability a customer will not join system

Average number of λ (1- PM)
customers in queue Lq = L -
µ

Average time customer L

spends in system W =
λ (1 - PM)

Average time customer 1

Wq = W -
spends in queue µ

Copyright 2006 John Wiley & Sons, Inc. 17-23

 Finite Queue: Example
First National Bank has waiting space for only 3 drive in window
cars. λ = 20, µ = 30, M = 4 cars (1 in service + 3 waiting)

Probability that
no cars are in 1 - λ /µ 1 - 20/30
P0 = = = 0.38
the system 1 - (λ /µ )M + 1 1 - (20/30)5

Probability of n=M 4
exactly 4 cars in λ 20
Pn = (P0) = (0.38) = 0.076
the system µ 30

Average number
λ /µ (M + 1)(λ /µ )M + 1
of cars in the L= - = 1.24
system 1 - λ /µ 1 - (λ /µ )M + 1

Copyright 2006 John Wiley & Sons, Inc. 17-24

 Finite Queue: Example (cont.)

Average number of λ (1- PM)

cars in the queue Lq = L - = 0.62
µ

Average time a car L

spends in the system W = = 0.067 hr
λ (1 - PM)

Average time a car 1

Wq = W - = 0.033 hr
spends in the queue µ

Copyright 2006 John Wiley & Sons, Inc. 17-25

 Finite Calling Population
Arrivals originate from a finite (countable) population
N = population size

1
Probability that no P0 =
N n
N! λ
customers are in system
Σ
n=0 (N - n)! µ

n
Probability of exactly n N! λ
customers in system Pn = P0 where n = 1, 2, ..., N
(N - n)! µ

Average number of λ +
Lq = N - (1- P0)
customers in queue µ
λ
Copyright 2006 John Wiley & Sons, Inc. 17-26
 Finite Calling Population (cont.)

Average number of
L = Lq + (1 - P0)
customers in system

Average time customer Lq

spends in queue Wq =
(N - L ) λ

Average time customer 1

W = Wq +
spends in system µ

Copyright 2006 John Wiley & Sons, Inc. 17-27

 Finite Calling Population:
Example
20 trucks which operate an average of 200 days before breaking
down (λ = 1/200 day = 0.005/day)
Mean repair time = 3.6 days (µ = 1/3.6 day = 0.2778/day)

Probability that no
P0 = 0.652
trucks are in the system
Average number of
Lq = 0.169
trucks in the queue
Average number of
L = 0.169 + (1 - 0.652) = .520
trucks in system
Average time truck Average time truck
spends in queue spends in system
Wq = 1.74 days W = 5.33 days

Copyright 2006 John Wiley & Sons, Inc. 17-28

 Basic Multiple-server Model
Two or more independent servers serve a single waiting line
Poisson arrivals, exponential service, infinite calling
population
sµ > λ
1
P0 =
n=s-1 n s

Σ
n=0
1
n!
λ
µ
1
+ s!
λ
µ
sµ
sµ - λ

Computing P0 can be time-consuming.

Tables can used to find P0 for selected values of ρ and s.
Copyright 2006 John Wiley & Sons, Inc. 17-29
 Basic Multiple-server
Model (cont.)
n
1 λ
Probability of exactly P0, for n > s
s! sn-s µ
n customers in the Pn = n
system 1 λ
P0, for n > s
n! µ

s
Probability an arriving 1 λ sµ
customer must wait Pw = P0
s! µ sµ -
λ
Average number of λµ (λ /µ )s λ
L = P0 +
customers in system (s - 1)!(sµ - λ )2 µ

Copyright 2006 John Wiley & Sons, Inc. 17-30

 Basic Multiple-server Model
(cont.)
Average time customer L
W =
spends in system λ

Average number of λ
Lq = L -
customers in queue µ

Average time customer 1 Lq

spends in queue Wq = W - =
µ λ

Utilization factor ρ = λ /sµ

Copyright 2006 John Wiley & Sons, Inc. 17-31

 Multiple-Server System:
Example
Student Health Service Waiting Room
λ = 10 students per hour
µ = 4 students per hour per service representative
s = 3 representatives
sµ = (3)(4) = 12

Probability no students
P0 = 0.045
are in the system

Number of students in
L = 6
the service area

Copyright 2006 John Wiley & Sons, Inc. 17-32

 Multiple-Server System:
Example (cont.)

Waiting time in the

W = L / λ = 0.60
service area

Number of students
Lq = L - λ /µ = 3.5
waiting to be served

Average time students

Wq = Lq/λ = 0.35 hours
will wait in line

Probability that a
Pw = 0.703
student must wait

Copyright 2006 John Wiley & Sons, Inc. 17-33

 Multiple-Server System:
Example (cont.)
 Add a 4th server to improve service
 Recompute operating characteristics
 P0 = 0.073 prob of no students
 L = 3.0 students
 W = 0.30 hour, 18 min in service
 Lq = 0.5 students waiting
 Wq = 0.05 hours, 3 min waiting, versus 21 earlier
 Pw = 0.31 prob that a student must wait

Copyright 2006 John Wiley & Sons, Inc. 17-34

Copyright 2006 John Wiley & Sons, Inc.
All rights reserved. Reproduction or translation
of this work beyond that permitted in section 117
of the 1976 United States Copyright Act without
express permission of the copyright owner is
unlawful. Request for further information should
be addressed to the Permission Department,
John Wiley & Sons, Inc. The purchaser may
make back-up copies for his/her own use only
and not for distribution or resale. The Publisher
assumes no responsibility for errors, omissions,
or damages
Copyright caused
2006 John Wiley by the use of these
& Sons, Inc. 17-35