Operations Research

Course Code: IPE 3103

Chapter 18
Queuing Systems

Instructor: Md. Rasel Sarkar

Dept. of IPE, RUET, Rajshahi, Bangladesh
Academic Year: 2020-21
Queuing System
 “Queue” is another name for a “waiting line”.
 Queueing theory is the study of the waiting line
systems.
 A waiting line system consists of two components:
1. The customer (people or objects)
2. The process or service system
 Whenever demand exceeds available capacity, a
waiting line or queue forms
Examples of Queueing System
 Jobs wait to be processed on a machine
 Planes circle in stack before given permission to
land
 Cars stop at traffic lights.
Elements of a Queueing System
 Customer: refers to anything that arrives at a facility
and requires service. e.g., people, jobs, airplane,
emails etc.
 Server: refers to any resource that provides the
requested service. e.g., repair persons, runways at
airport.
 Queue size : Characterized by the maximum
permissible number of customers that it can contain
may be finite (buffer area between two successive
machines) or infinite (mail-order facilities).
Setting an upper bound would on permissible
number be a complicating factor in the analysis. Thus
assumption of an infinite queue is the standard one
for most queueing models
 Queuing behavior:
Balking: When an arriving customer chooses not to
enter a queue because it’s already too long.
Reneging: When a customer already in queue gives
up and exits without being serviced.
Jockeying: When a customer switches between
alternate queues in an effort to reduce waiting time.
 State: The state of the system (n) denotes the
number of customers in the system.
Service System

 The arrangement of servers and waiting lines

 Queue discipline
 The arrival and service patterns
 Waiting Line Performance Measures
 The arrangement of servers and waiting lines
 Queue discipline
The order in which customers are selected from a
queue. The most common discipline is –
FIFO = First-in, first-out
LIFO = Last-in, first-out
SIRO = Service in random order
Highest profit customers first
Quickest service requirements first
Largest service requirements first
Earliest reservation first
Emergencies first Etc.
GD = General discipline (i.e., any type of discipline)
 The arrival and service patterns
Arrival rate:
 The average number of customers arriving per
time period
 Modeled using the Poisson distribution
 Arrival rate usually denoted by 
 Example:  = 50 customers/hour
i.e., 1/ = 0.02 hours between customer arrivals
Service rate:
 The average number of customers that can be
served during per time period
 Service times are usually modeled using the
exponential distribution
 Service rate usually denoted by µ
 Example: µ = 70 customers/hour
i.e., 1/µ = 0.014 hours per customer
Transient vs. Steady State
condition
The state of the system will be greatly affected by the
initial state and dependent on time during the early
operation of the system. The system is said to be in a
transient condition.

The steady state condition of a queuing system is

achieved after the system has been in operation for a
sufficiently long time and does not dependent on
initial state and time (time independent).
Terminology and Notation
c = number of servers
l= Arrival rate
µ = Service rate
n = State of system = number of customers in
queueing system.
L = length of Queue/ number of customers waiting for
service
= state of system - number of customers being served
=n-c
The following notation is used in a steady-state
condition:
 p = probability of exactly n customers in the system.
n
 L = average number of customer in the system

 Lq = average number of customer in the queue

 W = average waiting time in system

 Wq = average waiting time in the queue
Queueing system

L
Lq

l l
m
1
Wq 

W
Relationships between L, W, Lq, and Wq

Little’s Law
For any queuing system that has a steady state and
has an average arrival rate of ,
Average number of customer in system = arrival rate
x average waiting time
Ls = effWs
[rigorous proof is available of Little's formula]
Similarly, Lq = effWq
From figure –

These three relationships allow us to calculate all four

quantities Ls, Lq, W and Wq
he parameter λeff is the effective arrival rate at the system. It
equals the (nominal) arrival rate λ when all arriving
customers can join the system. Otherwise, if some customers
cannot join because the system is full (e.g., a parking lot),
then λeff < λ.
An arriving car may enter the parking lot at the rate λeff or it
may go elsewhere at the rate λlost. This means that
λ = λeff + λlost.
Waiting Line Performance Measures

 Ls = average number of customer in queuing

system

 Lq = average number of customer in the queue

 W: average waiting time in system

 Wq average waiting time in the queue

 Average number of busy servers, = Ls - Lq

 Facility utilization =
Pure Birth and Death Models
Birth refers to the only arrival of new customers
into the queueing system.
Death refers to only departure of served customers
from the system.
Pure Birth (only arrival) Model
To find the distribution of the number of arrivals during
a period t when the inter-arrival time is exponential with
mean 1/λ, let –
pn(t) = Probability of n arrivals during t
It can be shown that,

This is a poisson distribution with mean E{n} = λt.

The preceding result shows that if the time between
arrivals is exponential with mean 1/λ, then the number
of arrivals during a specific period t is Poisson with mean
λt. The converse is also true.
Pure Death (only departure) Model
In the pure death model, the system starts with N
customers at time 0, with no new arrivals allowed.
Departures occur at the rate m customers per unit
time. Let –
pn(t) = the probability of n customers
remaining after t time units
It can be shown that,

Thus it follows truncated poisson distribution.

The following table summarizes the relationships between
the exponential and the Poisson, given the arrival rate λ:
General Queueing Model/
Birth-and-Death Process
Goal: Given a queueing system with only birth and
death process the objective is to determine a closed
form expression of the probability of the number of
customer in the queue (pn) given only λ and µ. Also
determine L, W, Lq, and Wq

The birth–and-death process can be described as

where the state transitions are of only two types:
"births", which increase the state variable by one
and "deaths", which decrease the state by one, along
with the number of customer currently in the system.
Generalized model is based on the long-run or steady
state behavior of the queuing situation
Define,
λn= Arrival rate, given n customers in the system
µn= Departure rate, given n customers in the system
The probabilities Pn are determined by using the
transition-rate diagram in Figure below –

 For n > 0, state n can change only to two possible

states: n - 1 when a departure occurs at the rate λn, and
n + 1 when an arrival occurs at the rate µn.
 State 0 can only change to state 1 when an arrival
occurs at the rate λ0. Notice that µ0 is undefined
because no departures can occur if the system is empty.
Balance equation: Under steady-state conditions, for
n > 0, the rate at which the process enters state n
equals the rate at which it leaves state n, i.e.,
Mean entering rate = mean leaving rate
Now,
Expected entering rate into state n = λn – 1 pn - 1 + µn + 1pn + 1
Expected rate of leaving of state n = (λn + µn)pn

Equating the two rates, we get the following balance

equation:
λn – 1pn - 1 + µn + 1pn + 1 = (λn + µn)pn for n =1,2,3....
From transition rate diagram, the balance equation associated
with state 0, balancing equation is –

or,
Kendal Notation
A convenient notation for summarizing the characteristics
of the queuing situation is given by the following format:
(a/b/c): (d/e/f)

Here,
a = Arrivals distribution
b = Departures (service time) distribution
c = Number of parallel servers (= 1, 2, c, ∞)
d = Queue discipline
e = Maximum number allowed in the system
f = Size of the calling source (finite or infinite)
Common notations:
M – Poisson/ Exponential distribution
D - degenerate distribution (constant times)
G - General distribution (any arbitrary distribution
allowed)

Example: M/M/s model assumes that both inter-

arrival times and service times have an exponential
distribution and that the number of servers is s.
Specialized Queueing Model
based on Birth-and-death
Process
 Single-Server Model (M/M/1):(GD/ ∞/∞).
 Multiple-Server Model (M/M/c):(GD/∞/∞).
Single Server Model (M/M/1)
Goal: A closed form expression of the probability of the
number of jobs in the queue (Pi) given only l and m
Given:
• l: Arrival rate of jobs
• m: Service rate of the server
Solve:
• Ls: average number in queuing system
• Lq average number in the queue
• Ws: average waiting time in whole system
• Wq average waiting time in the queue
We have

Also, λeff = λ and λlost = 0, because all arriving

customers can join the system.
Generalized queueing model –

Letting, generalized model reduces to

To solve for p0 note that -

(geometric series, < 1)

<1
<1
Solving for Ls
Solving Ws, Wq and Lq
Multiple-Server Model (M/M/c)
This model deals with c identical parallel servers. The
arrival rate is λ and the service rate per server is µ.
Thus λn and µn are defined as –
Thus, for n < c

For n ≥ c

If
To solve for p0 note that -
Solving for Lq
Solving Ws, Wq and Ls