Queueing Theory

Introduction
In this chapter we will study a class of models in which customers arrive in some random
manner at a service facility. Upon arrival they are made to wait in the queue until it is their turn
to be served. In our daily life, there are many reasons to form the queue. For example, Persons
waiting in hospital for treatment, A person waiting in a queue to purchase a ticket in either a
bus station or a railway station, a person waiting near ticket counter of a cinema hall, a person
in queue in a cash counter for withdrawal of money, a machine repaired and waiting for the
attention of a maintenance crew is an element or a customer, etc. The main objective of
Queueing theory is to provide a reasonably acceptable service to waiting customers. Queueing
theory regulates the measures of performance of waiting lines, such as the average waiting time
in queue and the productivity of the service facility, which can then be used to design the service
installation. They are 2 things in Queueing theory, one is service and also the different one is
client or component. Here client or component represents someone or machine or the other
issue, that is in want of some service from sexual union purpose. Service represents any variety
of attention to the client to satisfy his want. Customers suggests that “a person reaching to
hospital”, Service suggests that “Doctor provides a treatment to a person”. A queue is outlined
as “A cluster of customers / things waiting at some place to receive attention / service as well
as those receiving the service.”

We can see here that each one the client can reach and waits to avail the service at station. once
the station has no desired capability to serve all at a time the client must look ahead to his/its
flip ensuing the formulation of a line of shoppers that is mostly called a queue. generally we
are able to say that a flow of shoppers. Queue may be decreased in size by providing extra
service facilities. Sometimes it may reduce the profit of the company. On the other hand long
queues may reduce the sales and lost customers. Therefore the aim of Queueing theory is to
balance between the costs associated with prevention of waiting in order to maximize the profit.

In general, the Queueing system contains of one or more queues and one or more servers, and
operates under a set of procedures. If the system is free at the time of arrival of a customers,
the customers can directly enter in to the counter for receiving service and then leave the
system. The study of queues deals with quantifying the phenomenon of waiting in lines using
representative measures of performance, such as average queue length, average waiting time
in queue, and average facility utilization. The following example demonstrates how these
measures can be used to design a service facility.

Example Members of queue Server(s)

Hospital Patient Doctors
Railway station Passenger Ticket counter
Toll gates Vehicle Toll collectors
Bank Account holder Bank cashier
Traffic signals Vehicles Traffic points
Airport Aero planes Runway signals

In general, we are able to say that a flow of shoppers from infinite or finite population towards
the service facility forms a queue or queue on account of lack of capability to serve all at a
time. The higher than discussion clarifies that the term client we have a tendency to mean to
the arrival unit that needs some service to be performed at the station. Queues or waiting lines
stands for variety of shoppers waiting to be maintained. Queue doesn't embody the client being
maintained. the method or system that performs the services to the client is termed as service
channel or service facility. so from the higher than we have a tendency to see that waiting lines
or not solely the lines shaped by men however additionally the opposite things like railway
coaches, vehicles, material etc.

Figure . 1

HISTORICAL DEVELOPMENT OF THE THEORY

During 1905, Erlang, a Danish engineer has begins work on queueing theory. telephone unit
type a Mathematical modelling to work out the consequences of instable service demand on
the utilization of automatic dialing kit. it's been solely since the tip of warfare II that job on
queue models has been extended to a different types of issues. In 1927, Millins are established
the idea any then by William Thornton and Fry. however Edward Calvin Kendall has set a
scientific and mathematical approach to queue downside in 1951. once 1951 important work
has been exhausted queue theory, therefore on modify it to use to sorts of issues stumble upon
in industries and society. In toady’s situation a large style of apparently various issues things
square measure recognized as being represented by the overall queue model. In any queuing
system, we've associate input that arrives at some facility for service or process and therefore
the time between the arrivals of individual inputs at the service facility is usually random in
nature. Similarly, the time for service or process is usually a stochastic variable. One best
example of this might be quoted because the management of waiting time of a client in queue
advanced of Tirupathi Temple. the current system of ligature a belt with time to the hands of a
client is that the results of application of queuing theory. Another example is computerised
reservation of rail journey.
Uses of Queuing theory:
Queuing theory is used to identify and correct points of congestion in a process. The queue
may consist of people, things, or information. In any case, they are being forced to wait for
service. That is inefficient, bad for business, and annoying (when the queue consists of
people).
Queuing theory is used to analyze the existing process and map out alternatives with a better
result. Queuing theory is also applied to moving objects or information through a “line.” For
example, an auto manufacturer may look to queuing theory for guidance on the most efficient
way to set up its assembly lines. A transport company, such as FedEx or UPS, can use
queuing theory to determine the most operationally efficient manner of transferring packages
from one transport vehicle to another.
In the world of business, queuing theory can help a company’s executives determine the best
way to set up and organize business operations so as to maximize both sales and customer
service satisfaction.
Financial analysts may construct models based on queuing theory to make projections about
how changing an operational variable may improve queuing efficiency and, as a result,
bottom-line profitability. The use of queuing theory has become so popular that there are
now online queuing calculators available that can do a basic analysis of a given queuing
setup.

Characteristics of Queueing theory

There are, In general, six characteristics of Queueing theory

1. Arrival pattern of a customers
2. Service pattern of a servers
3. Queue discipline
4. System capacity

Figure. 2 Components of Queueing system

Arrival (or input) pattern of a customer

The input defines the way in which the customers arrives and join in the system. The system
includes both queue and the server. The customer arrivals is random in nature, which cannot
be predicted. Servers cannot control the arrival of a customers. Therefore the number of arrivals
in a particular time period cannot be treated as a constant. The arrival of your time between one
clients to future customer is thought as inhume time of arrival (i.e. the time between 2
succeeding arrivals). we are able to see at gas bunks or price tag counters or any such service
facility that the client arrives willy-nilly individual or in batches. the scale of client arrival
absolutely supported the scale of the population. The population could also be finite or infinite.
The number of arrivals per unit time (commonly used) has a poisson distribution with
1
mean 𝜆. Whereas, inter arrival time follows an exponential distribution with mean 𝜆.
Moreover the input pattern should state the size of queue requiring for the service (finite or
infinite), the number of queues that are permitted to form.

The size of the queue or the customer waiting time is generally depends on the conduct of the
consumer. Here the conduct refers to the impatience of a customer during the stay in the line.
Customer behavior can be classified as balking, reneging, collusion and jockeying. Balking
refers that the customer ma not join in to the queue after seeing the long length of it. This
performance may effect in losing a customer by the organization. Reneging refers the customer
joins the queue and after waiting for certain time loses his patience and leaves the queue. This
behaviour of the customer may also cause loss of customer to the organization. Another
behavior of the customer is collusion. In this case one customer may represent the group of
customer. Here size of the queue may be small but time of service for an individual customer
will be more. Jockeying refers that the number of waiting lines depending on the number of
service stations. A client in one in every of the queue when seeing the opposite queue length,
that is shorter, with a hope of obtaining the service, might leave this queue and be a part of the
shorter queue. maybe true could also be that alternative queue that is shorter could also be
having a lot of range of Collaborated customers. In such case the likelihood of obtaining service
to the client WHO has modified the queue could also be terribly less.
There are three states of nature in case of arrivals in a queue system (i) steady state (ii) transient
state, and (iii) explosive state. The input and output parameters are independent of time or the
system is independent of time, the system is said to be in steady state. Queuing theory analysis
involves the study of a system’s behaviour over time. A system is claimed to be in transient
state. in a very state of affairs, wherever arrival rate of the system is larger than its service rate,
a gradual state can not be reached in spite of the length of the time period. Here queue length
can increase with time and on paper it may build up to time. Such case is termed the explosive
state.

Figure. 3 Characteristics of input Pattern

Service pattern (or) service mechanism

It represents the pattern in which the number of customers lives the system. Departures may
also be represented by the service time which is the time period between two successive
services. The number of customer served per unit time is called the inter service rate. The
number of customers serviced per unit time follows a poisson distribution with mean 𝜇 or the
1
inter service time has an exponential distribution with mean μ. The time required to serve the
customer is not a constant.
There are two types of server (i) single server (ii) multiple server.
(i) Single server: If the service suppliers has provided single server to the purchasers, just one
unit will be served at a time. Customers may form a single queue and get serviced. For eg.
Petrol bunks, the vehicle enters the petrol station. If there is only one petrol pump is there, it
joins the queue near the pump and when the term comes, get the fuel filled and soon after leaves
the queue.
(ii) Multiple server: If a service providers has provided a multiple server to the customers, more
than one queue may formed and more than one customer can be served at a time. Note that the
system can also have a number of service channels where the customer may be arranged in
parallel or in series. If a patient enters the hospital, get examined by doctor and take prescription
goes to medical shop to takes medicine and then goes to nurse have the injection and leaves the
hospital.
 Figure 4. Single server Queueing model

Figure 5. Multiple server Queueing model (parallel)

Figure 6. Multiple server Queueing model (series)

Queue discipline (or) Service discipline
The queue discipline (or) service discipline specifies the manner in which the customer from
the queue are equal and led the manner in which they are selected for service. There are
many ways in which consumer called to serve.
(i) First In First out (FIFO) is commonly followed by all the organizations. First Come First
serve(FCFS) or first in first out is followed, which the customer service served in strict order
of the arrival and it’s understand that it gives a good morale and discipline in the queue.
(ii) Last in first out (LIFO) or Last Come First Served (LCFS) is the customer arrived last to
get chance served very first by the server. This queue discipline is commonly followed in
inventory system. For example, the present stock in a factory will kept in the down but the
new ordered material will stock on the top. While consuming the top material (last in) is
being consumed.
(iii) Service In Random Order (SIRO) is the customer may be first or last in the queue does not
a problem, the servicing facility calls the customer in random order without considering the
order of arrival.
(iv) Service By Priority disciplines area unit those wherever any arrival is chosen for service
prior to another customers already in queue. Some times within the hospital this service
discipline could follow. as an example, we will quote that in a very doctor’s search, once the
doctor is treating a patient with abdomen pain, suddenly a patient with heart stroke enters the
doctors search, the doctor asks the patient with abdomen pain to attend for a few time and
provides attention to heart patient.
System capacity
Maximum range of shoppers allowed within the system is either infinite or finite. It in some
facility is just a restricted range of shoppers allowed within the system however generally
range of shoppers within the system is time.

Figure. Service Mechanisms

Classification of Queueing model

Normally queuing models are entirely specified in the following simple form (a/b/c) : (d/e)
(I.e. Poisson Arrival / Poisson output / Number of channels / Infinite capacity / FIFO Model)
where ‘a’ denotes probability law for the arrival (or) probability distribution type of arrival per
unit time, ‘b’ is type distribution of service per unit time (or) probability law according to which
the customer are being served, ‘c’ is the number of servers (single or multiple servers), ‘d’
denotes the maximum number customers allowed in the system (finite or infinite ) and ‘e’ is
the queue discipline.
The above notation is called Kendal’s notation
The first four models can be classified as
1. (M / M /1) : ( / FIFO) - Single server infinite capacity
2. (M / M / S ) : ( / FIFO) - Multiple server infinite capacity
3. (M / M /1) : (k / FIFO) - Single server finite capacity
4. (M / M / S ) : (k / FIFO) - Multiple server finite capacity

Where ‘M’ stands for Markov (Both input and output pattern in unit time t follows the
Poisson process which is a continuous time Markov chain)
Poisson Arrival / Poisson output / Number of channels / Infinite capacity /FIFO
Model, (M/M/1):(∞/𝐅𝐈𝐅𝐎):
Formulae used

1. Average number of arrivals per unit of time = 𝛌

2. Average number of units served per unit of time = 𝛍

𝛌
3. Traffic intensity or utility ratio = 𝛒 = the condition is: (𝛍 > 𝛌)
𝛍

4. Probability that the system is empty= 𝐏𝟎 = (𝟏 − 𝛒)

5. Probability that there are ‘𝐧’. units in the system= 𝐏𝐧 = 𝛒𝐧 𝐏𝟎
𝛒 𝛌
6. Average number of customers in the system (𝐋𝐒 ) = or (𝛍−𝛌)
(𝟏−𝛒)

𝛒𝟐 𝛌𝟐
7. Average number of customers in the Queue( 𝐋𝐪 ) = =
(𝟏−𝛒) 𝛍(𝛍−𝛌)

𝐋𝐒 𝟏
8. Average waiting time of customer in system ( 𝐖𝐒 ) = =
𝛌 𝛍−𝛌
𝐋𝐪 𝛌
9. Average waiting time of customer in Queue ( 𝐖𝐪 ) = =
𝛌 𝛍(𝛍−𝛌)
𝛍
10. Average number of customers in non-empty space ( 𝐋𝐰 ) =
(𝛍−𝛌)

11. 𝐏(𝐰 > 𝟎) = System is busy = 𝛒

12. Probability that the no of customers in the system exceeds K. P (N > K) = 𝛒𝐊+𝟏

13. Probability intensity function of waiting time in the system = (𝛍 − 𝛌)𝐞−(𝛍−𝛌)𝐰

14. Probability that waiting time of customer exceeds in system P (WS > t) = 𝐞−(𝛍−𝛌)𝐭
15. The Average waiting time of the customer in the queue if he has to wait
𝟏
E (Wq / Wq > 0) =
𝛍−𝛌
Problems:

1.A T.V. fixer finds that the time spent on his jobs have an exponential distribution with
mean of half-hour. If he repairs sets in the order in which they come in, and if the arrival of
sets is approximately Poisson with an average rate of 10 per 8hour day, what is repairman’s
expected idle time every day? How many jobs are ahead of the average set just brought in?
Solution:
This problem is Poisson arrival/Negative exponential service / single channel
/infinite capacity / FIFO type problem.
Data: λ = 10 sets per 8-hour day
= 10 / 8 = 5/4 sets per hour.
μ = (1/30) × 60 = 2 sets per hour.
λ
Hence, Utility ratio = ρ = μ = (5/4) / 2 = 5 / 8 = 0.625.
This means out of 8 hours 5 hours the system is busy i.e. fixer is busy.
Probability that there is no queue = The system is idle =1- ρ = 1- (5 / 8) = 3 / 8
= That is out of 8 hours the fixer will be idle for 3 hours.
Number of sets ahead of the set just entered = Average number of sets in system
λ ρ
= μ−λ = (1−ρ)
= 0.625 / (1 - 0.625)
= 5 / 3 ahead of jobs just came in.

2. A branch of a Nationalized bank has only one typist. Since typing work varies in length
(number of pages to be typed), the typing rate is randomly distributed approximating a Poisson
distribution with a mean service rate of 8 letters per hour. The letter arrives at a rate of 5 per
hour during the entire 8- hour workday. If the typist is valued at Rs. 1.50 per hour, determine:
(a) Equipment utilization, (b) The percent time an arriving letter has to wait, (c) Average
system time, and d) Average idle time cost of the typewriter per day.

Solution:

Data = arrival rate = λ = 5, Service rate μ = 8 per hour.

λ
Hence ρ = = 5 / 8 = 0.625
μ
(a) Equipment utilization = Utility ratio = ρ = 0.625,
i.e. 62.5 percent of 8hour day the equipment is engaged.
(b) Percent time that an arriving letter has to wait = As the machine is busy for 62.5 % of the
day, the arriving letter has to wait for 62.5 % of the time.

(c) Average system time = Expected (average) a customer spends in the system
1
= μ−λ
1
= 8−5
= 1/3 hour
= 20 minutes.
d) Average idle time cost of the typewriter per day
= 8 hours × idle time × idle time cost
 = 8 × (1 -5 / 8) × Rs. 1.50
= Rs. 4.50.

3.Arrivals at a telephone booth are considered to be Poisson with an average time of

12 min. between one arrival and the next. The length of a phone call is assumed to be
distributed exponentially with mean 4 min.
a) Find the average number of persons waiting in the system.
b) What is the Probability that a person arriving at the booth will have to wait in the
queue?
c) What is the probability that it will take him more than 10 min. altogether to wait for
the phone and complete his call?
d) Estimate the fraction of the day when the phone will be in use.
e) The telephone department will install a second booth, when convinced that an arrival
has to wait on the average for at least 3 min. for phone. By how much the flow of arrivals
should increase in order to justify a second booth?
f) What is the average length of the queue that forms from time to time?
Solution:
1
Mean inter-arrival time = 𝜆 = 12 min.
1
Therefore, mean arrival time = 𝜆= 12 per minute.
1
Mean inter-service time = = 4 min.
𝜇
1
Therefore, mean arrival time = 𝜇= 4 per minute.
a) Average number of persons waiting in the system

λ
= μ−λ
1
12
= 1 1
( − )
4 12
= 0.5 customer.

b) P (W > 0) = 1- P (W = 0)
= 1- P (no customer in the system)
= 1- P0
λ
= 1- (1 - μ )
1
λ 12 1
= = 1 =
μ 3
4

c) P (W > 10) = 𝑒 −(𝜇−𝜆)10

1 1
−( − )10
=𝑒 4 12
5
= 𝑒 − ( 3 ) = 0.1889

d) P (the phone will be idle) = P (N = 0) = P0

λ 2
= (1 − μ ) = 3
2 1
Therefore P (the phone will be use) = 1− 3 = 3
 1
Or the fraction of the day when the phone will be in use = 3.

e) The second phone will be installed, if E ( Wq ) > 3.

λ
If >3
μ(μ−λ)
𝜆𝑅
1 1 > 3,
( − 𝜆𝑅 )
4 4
Where 𝜆𝑅 is the required arrival rate
3 1
If 𝜆𝑅 > 4 ( 4 − 𝜆𝑅 )
3
If 𝜆𝑅 > 28
3 1 1
Hence the arrival rate should increase by − = per minute, to justify a second
28 12 42
phone.
f) Average length of the queue that forms from time to time
μ
=
(μ−λ)
1
4
= 1 1
( 4 − 12 )
= 1.5 persons.

4. There is a congestion of the platform of a railway station. The trains arrive at the rate of 30
trains per day. The waiting time for any train to hump is exponentially distributed with an
average of 36minutes. Calculate:
(a) The mean queue size,
(b) The probability that the queue size exceeds 9.
Solution
Data: Arrival rate 30 trains per day, service time = 36 minutes.
λ = 30 trains per day.
1
Hence inter arrival time = 30(24 Χ 60 )= 48 minutes.
1
λ = 48 minutes
1
μ = 36 minutes.
λ 36
ρ= = 48 = 0.75
μ
ρ
(a) The mean queue size = E (n) =
(1−ρ)
= 0.75 / (1 - 0.75) = 0.75 / 0.25
= 3 trains.
(b) Probability that queue size exceeds 9 = Probability of queue size ≥ 10
= 1 - Probability of queue size less than 10
= 1 - (P0 +P1 +P2 ...+ P9)
= P0 (1+ ρ + ρ2 + .... ρ9)
= 1 - {(1- ρ) [(1- ρ10) / (1- ρ)]}
= 1 - (1- ρ10)
= ρ10 = (0.75)10
= 0.06 approximately).