Queuing (or Waiting Line)

Theory
8-1. INTRODUCTION
Queues or waiting ines are very common in everyday life. For example,
customers wait for service in front of railway booking offices, post offices, bank
counters, doctor clinics, bus stops, etc. A queuing problem arises when the current
service rate of a facility falls short of the current rate of customers' demand. The
word queue refers to customers waiting for service in a line. It may be a physical line
of customers or a waiting list for berth on
a train, etc.
Some common examples of waiting lines are given in the following table:
WAITING LINE OR QUEUING SITUATIONS
Example Unit arriuing Service or facility Process performed
required
Manually paced Parts to be assembled Assembly line or Assembling
assembly lhne assembly operations
Inventory of items in a Order for withdrawal Warehouse Replenishment of
warehouse inventory
Tools in the tool room Tools for service Tool room Servicing the tools
Ships entering a port Ships Docks Loading and unloading
Maintenance and repair Machine break-down Repair crew Repair work
of machines
Customers at checking Customers Checking counters Customs clearance
counters at airports
Buses at bus terminal Buses Bus terminal Parking
Purchase of groceries in Customers loaded with
super bazar
Check out counter Billing and payment
grocery

Sometimes the operations manager is faced with the decision of determining just
how many service faciljties should be in operation in order to adequately meet
customer demand. Examples of this type of situation include determining the
appropriate number of checkout clerks in a supermarket, tellers in a bank,. doctors
on duty in an emergency room, gas pumps in a service station and so on. Queuing
theory is a quantitative technique useful for determining the optimumnumber of
service facilities.
There are two basic types of costs associated with waiting-line problems.
First, there are the fairly tangible' costs involved in operating each service facility
like the costs for equipment, materials, labour, etc. These costs, of course, rise as.
the number of service facilities put into operation increases. On the other hard
there are the relatively intangible' costs associated with causing customers Yo have
to wait in line for some period of time prior to being waited upon-physical
discomfort, adverse emotional reactions, reduced or lost sales and so on. Of course,
 the number of
as
service facilities in Cost (Rs.)
operation increases, the time the
customer has to wait in line, on the
average, decreases, and hence s0 Total expected
too do these costs. -cost (TC)
As shown in the adjoining fig.
8.1, the total of these two basic TCMIN Cost of providing
service
types of costs goes to a minimum
at some specific number of failities. Cost of waiting
This then is the optimum number of time
service facilities which should be
scheduled by the operations QoPT
manager-optimum because it Service facilities in Operation
minimizes the total cost of both
operation service facilities and Ag. 8-1:Costs for the General Waiting Line situation
waiting to be serviced at them.

Queuing modelsare appliedto problemsconcemingthearivalof customerstorecelve

some service at a collection ofservice.points. The objective of these models is to
detemine an optimum number ofpersonnelorfacilities to service customers, balaincing the
cost of service with the cost of waiting or congestion.

Informational requirements. For a queuing model, information regarding the

following components is required
1. Size of the calling population which may be finite or infinite.
2. Arrival size can be single or a batch.
3. Arrival distribution. The pattern or distribution of arrivals can be constant or random.
4.
4. Size of the queue which may be finite or infinite.
5. Attitude of the customers. Customers can be either impatient (balking or reneging)
or patient (voluntary or involuntary).
6. Queue discipline. It refers to the decision rule by which the customers are selected
from queue for service.
7. Service distribution. The distribution of service time can be constant or random.
8. Configuration of the service facilities. We can classify service facilities in terms of
their configuration of channels (single or multiple) and phases (single or multiple).
9. Queuing system. The system consisting of arrival of customers, waiting in queue,
picked up for service according to a certain discipline, being serviced and the departure
of customers.

8-2. FLEMENTS OF THE QUEUING SYSTEM

The basic elements ofthe queuing system are as follows:
1. The Input Process. This element is concerned with the pattern in which the
customers arrive and join the system. An input source is characterised by:
(a) its size
(b) arrival time distribution of customers.
C) mean time between intervals
(d) attitude of the customers.
In terms of size, an (population) is finite or infinite. If the rate at
input source
which the source generates the customers is affected by the number of customers
in the service system (customers in queue and those in service) population is
considered finite, otherwise it. is intinite.
 The periods between the arrival of individual customers may be constant or
scattered,in some tashion. For example,. in restaurant the arrival times of customers
are distributed randomly. But in a clinic appointments can be so given that patients
arrive at specified intervals of time.
Mean arriual rate refers to the average number of customers arriving per unit of
time. It is often represented by Greek letter 2 known as 'Lambda'. Mean time
between arrivals is represented by 1/^.
Attitude or behaviour of customers joining the service system may be of several
types. If customers stay in the system until served, they are called 'patient'
customers otherwise they are known as 'impatient' customers. Impatient customers
may wait tor some time and leave the system if not served or they move from one
waiting line to another.
2. with service
and
Service mechanism. The service mechanism concerned
the
is time
service tacilities. Service time is time interval from the commencement of
service to the completion of service. It may either be fixed or distributed in
accordance with some probability distribution. Service facilities or service channels
may be in a sequence or parallel. In the former case a customer must go through
one facility after another in a particular sequence. In the latter case customers are
serviced in no particular sequence, e.g., parking of vehicles. Service facilities can be
of the following types
(a) Single channel facility, i.e., one queue-one service station facility wherein the
customer waits till the service point is ready to take him for servicing.
(b) One queue-several service station facilities wherein the customers wait in a
single queue until one of the service stations is ready to take them for servicing.
(c) Several queue-one service station wherein there are several queues and the
customer can join any one of these but there is only one service station.
(d) Multi-channel facility where there are many service station failities and
many queues.
(e) Multi-stage channel facilities wherein several service stations are there each
providing a
specialised service. Customers pass through each of the service stations
before leaving the system.
With respect to service mechanism the following information is considered
relevant
() Average number of customers being served in one unit of time at a service
station. This is often represented by Greek letter .
it) Average time taken to service a customer (represented by 1/p).
3. Queue discipline. Queue discipline refers to the order in which the service
station selects the customers from the waiting line. Customers may be taken into
service in order of their arrival. This is called the first come first served' service
descipline. Sometimes customers may be given a priority basis for service or may be
selected at random. Another aspect of queue discipline is whether a customer in a
queue can move to a shorter queue in the multi-channel system. Queue discipline also
refers to the manner in which the customers form into queue and the manner in
which they behave while being in the queue.

4. Output of the Queue. Generally this tactor is not important but in some cases
it may influence the service/arrival rate. For example, if there is only one dopr to
service point through which people enter and leave after being served, it is possible
that people leaving could affect the rate of arrival. In a single channel facility the
output of the queue does not pose any problem because the customer departs after
receiving the service. But the output of the queue becomes important. When the
System is of multistage channel facilities due to the possibility of a service station
breaking down can have repercussions on the queues. The line before the
breakdown will lengthen and the line following the breakdown will diminish.
Ihus, a queuing model can be described as composed of customers arriving for
Service,waiting for service and then leaving the system after being served. It
essentially involves the study of behaviour of waiting lines with the help of
mathematical techniques utilising the concept of stochastic procesS.
8:2-1. Operating Characteristics of Queuing System.The operating
characteristics of a queuing system refer to the numerical values of the probability
distributions of various decision variables like arrival rate, number of facilities, service
time, line length, priority system, etc. Some common characteristics are given
below:
1ueue length. Probability distribution of queue length can be obtained with the
help
o f the given probability distribution of the arrival and service process. A large
queue indicates poor service facility or a need for more space. On the other hand,
small qyeue indicates excess of service facilities.
2 . Waiting time in queue. It refers to the time spent by a custonmer in the queue
before the commencement of his service. Long waiting times may increase the
customer's dissatisfaction and potential loss of future revenues.
3 aiting time in system. This is the total time spent by a customer in the
quede plusservice time. Long waiting time may indicate need for a change in priority
rules.
A queuing system is said to be in transient state when its operating
characteristics are dependent upon time. If the operating characteristics become
independent of time, the queuing system is said to be in a steady state.

8-3TYPES OF QUEUING MODELS

There are several types of queuing models. Some of them are:
1. Single queue-single service point
2. Multiple queues-multiple service points
3. Simple Queues-multiple service points
4. Multiple queues-single service point.
The most common case of queuing models is the single channel waiting line.
8-4. MANAGERIAL APPLICATIONS OF QUEUING THEORY
Queuing theory is a valuable tool for business decision-making. It can be applied
to a wide variety of situations for scheduling. Some of these are given below:
(a) mechanical transport fleet
(b) scarce defence equipment
c) issue and return of tools from tool cribs in plants
(d aircrafts at landing and take-off from busy airports
(e) jobs in production control
parts and components in assembly lines
routing sales persons
(h) inventory analysis and control
replacement of capital assets
minimisation of congestion due to traffic delays at booths.
Queuing theory has generally been applied by factories, transport companies,
telephone exchanges, computer centres, retail stores, cinema houses, restaurants,
banks, insurance companies, traffic control authorities, hospitals, etc.
 $-5. AbVANTAGES AND LIMITATIONS OF QUEUING THEORY
VIt offers the following benefits :
1. Queuing theory provides models that are capable of determining arrival
pattern of customers or most appropriate number of service stations.
2. Queuing models are helpful in creating balance between the two opportunity
costs for optimization of waiting costs and service costs.
3. Queuing theory provides better understanding of waiting lines so as to
develop adequate service with tolerable waiting.
Major limitations of queuing theory are
1. Most of the queuing models are very complex and cannot be easily
understood. The element of uncertainty is there in almost all queuing situations.
Uncertainty arises due to:
(i) We may not know the form of theoretical probability distribution which
applies.
(i) We might not know the parameters of the proces even when the
particular distribution is known.
(ii) We would simply be knowing only the probability distribution of out-
comes and not the distribution of actual outcomes even when (i) and (ii)
are known.
2. In addition to the above complications, queue discipline may also impose
certain limitations. If the assumption of 'First come first served' is not a true one
(and this happens in many cases) queuing analysis becomes more complex.
3. In many cases, the observed distributions of service times and time between
arivals cannot be fitted in the mathematical distributions usually assumed in queuing
models. For example, the Poisson distribution which is generally supposed to apply
may not fit many business situations.
4 In multi-channel queuing, the departure from one queue often forms the
arrival of another. This makes the analysis more difficult.

8-6,SYMBOLS AND NOTATIONS

The following symbols and notations are generally used in queuing models
n total number of customers in the system, both waiting and in service
= average number of customers arriving per unit of time
H average number of customers being serviced per unit of time
C = number of parallel service channels
Ls or Eln) average number of customers in the system, both waiting and in
service
Em) = average number of customers waiting in the queue
La
W or
or
E(v) = average waiting time of a customer in the system both waiting and in
service

Wg or E{w) average waiting time of a customer in the queue

=

PC)= Probability that there are n customers in the system at any time t,
both waiting and in service
p Trafic intensity or utilization factor which represents the proportion

oftime the servers are busy =

8-7. SINGLE CHANNEL
QUEUING MODEL
he most common case of queuing problems involves the single channel or
single server, waiting line.
The following assumptions are usually made in this type of system:
.There is only one type of queue disclpline, t.e., first come first served.
2. There are steady state (stable) conditions, t.e., the probability that n items are in
queue at any time remains the same with the passage of time.
Mean service rate exceeds the mean arrival rate.
4. Arrival rates follow Poisson distribution and service times follow exponential
distribution.
5. Both the number of items in queue at any time and the waiting line experdenced by aa
particular item are random variables.
6. Arrivals are from infinite population.
7. Arrivals are independent of each other, i.e., what happened in previous time periods is
not relevant to the future time periods. Similarly services are also independent of each
other.
8. The waiting space available for customers in the queue is infinite.

Service Departure
-Arrivals OOOD saneeeeeoe
facility
sooeeeaeeeee
after service
System

Fig.8-2. Single-Channel Waiting Line

Then the model is given by the following equations:
Parameters System Queue Remarks
1. Average waiting time per
unit W -1 W -)**
Average length (Number of
2.
units) L-
3 Non-empty queue length

4. Probability that there will

be n or more units in the
system
P,-Pox(
5. Probability that waiting e-Jt
Po1
time is more than t.

Remarks. 1. Unless specified, the average waiting time means waiting time in queue (not in system).
2. The above relations hod good for Poisson/exponetial distribution with A and H as arrival and
service rates.
3. Utilization factor for the system, p - Nu
88. MULTIPLE CHANNEL QUEUING MODEL
In a multiple-channel queuing system, two or more servers or channels are
available to handle customer who arrive for service.

Server Departure
-Arrivals- O+Server >after
Queue
Server service

Fig.8-3. Multiple-Channel Queuing System

With the number of calling units
represented by n and the number of channels of
service stations by k, the descriptive characteristics of the
multi-channel system are
summarized as follows:
=
average arrival rate
average service rate at each channel
1. The probability that there are no customers units in the
or
system
1
1-(/kp)
2. The probability that there are n units in the system :

Po. if n <k
P
(/4
kI kn-F Po, if n 2 k
3. P n2k) =
Probability that a custormer has to wait
H
(k-1)!(ku-O
4. Average (expected) number of customers in the system

k-1)! (ku-312 Po +
5. Average queue length (i.e., expected number of customers waiting in the
queue

=L,-:
k-1)! (ku-2 Po
6.
Averagetime an arrival spends in the system:
W, =
(k - u-P Po* APot
 . Average waiting time of an arrival (1.e., expected time a customer spends in
the queue waiting for service)
Wa (k-1)! (ku - PoW-
8 Utilization rate :
P k