Int. J. Mathematics in Operational Research, Vol. 12, No.

1, 2018 1

Passengers queue analysis in international airports

terminals in Kerala using multiphase queuing system

Roshli Aniyeri*
Department of Mathematics,
MGGA College,
New Mahe, Thalassery,
Kerala 673311, India
Email: purerose2007@rediffmail.com
*Corresponding author

Ratnam Nadar
Noorul Islam University,
Kanyakumari District, India
Email: crnadar@gmail.com

Abstract: Waiting period is a global problem that almost everyone has to face,
which causes a great waste of time for everyone. It is well known that all these
waiting line problems critically restrict further development. The focus
of this study is to deal with passengers’ queue issues of the international
airport terminals of Kerala. Queuing theory is a mathematical approach to
the study of waiting period in queues. This study evaluates the effectiveness of
multi-server queuing model. The multi server approach of modelling was
adopted in this cram to develop a mathematical model to solve problem of
queuing of air transport passengers at the international airports in Kerala. The
airport in the aviation industry of the country faces problems of many
passengers queuing for boarding, departure with different arrival rate due to
non availability of state of the art logistics management mechanisms for
predicting the nature and service demands of travellers. The passengers’
average wait time for reaching the gate area measures system performance. A
mathematical queuing model was developed in this study and comparisons are
made using analysis of variance (ANOVA).

Keywords: multi server queuing modelling; airport passengers; queuing steady

state condition; estimated queue length; utilisation factor; analysis of variance;
ANOVA.

Reference to this paper should be made as follows: Aniyeri, R. and Nadar, R.

(2018) ‘Passengers queue analysis in international airports terminals in
Kerala using multiphase queuing system’, Int. J. Mathematics in Operational
Research, Vol. 12, No. 1, pp.1–30.

Biographical notes: Roshli Aniyeri is a PhD student in Queuing Theory at

Noorul Islam University Kanyakumari, India. She holds a Master’s and a
Bachelor’s degree in Mathematics at Calicut University. She has also done her
MPhil in Mathematics at Bharathiyar University. Her research interests include
queuing theory, simulation and Anova.

Copyright © 2018 Inderscience Enterprises Ltd.

2 R. Aniyeri and R. Nadar

Ratnam Nadar, previously Director of P.G. Department of Mathematics, is

currently a Dean of Students’ Welfare of Noorul Islam University. He has done
his PhD in Madras University. His main research areas are queuing theory, time
series analysis, system analysis, risk and uncertainty.

1 Introduction

The Indian aviation industry is one of the fastest growing airline industries in the world.
The history of Indian aviation industry started in December 1912 with its first domestic
air route between Karachi and Delhi. The airport is the gateway of any country. A tree
cannot make a forest, so says an adage. Thus the economy of a nation depends on some
sectors like the agricultural sector, the health sector, the works and housing sector and
most importantly, the airport, banking sector among others. All these sectors work
together and integrate to make the life of a nation. A breakdown in any of these sectors
will definitely have untold effects on the progress and success of the nation. The airline
industry is important for the global economy. Airports, in particular, hub airports are the
backbone of air transportation. Transportation plays a vital role in the changing global
economy, linking people and places, facilitating trade and tourism, and encouraging
economics competition and specialisation.

1.1 Problem definition

There are three international airports in Kerala. Therefore it is decided to collect data
from three international airports in Kerala for this research and compare it. The main
purpose of this paper is to review the application of queuing theory and to evaluate the
parameters involved in the service unit of the airport terminals in Kerala. The broad
objective of this study is to gain an understanding of the application of queuing theory to
the problem of airport terminals in Kerala.
The ultimate aim is to build models of better quality and to understand the key
concepts and to propose to develop a queuing model of the airports.

1.2 Problem statement

This study deals with the problem of finding the airport congestion problem in
international airports in Kerala. For a better understanding of airport capacity and
congestion problems, queuing theory gives great insights. As de Newville and Odoni
(2003, pp.819–863) point out, all airport facilities can be described as systems of queuing
systems, where arrivals at a service facility are randomly distributed, waiting lines form
and users are therefore delayed and have to wait before being served. Queues develop at
airports in numerous contexts (ticketing, baggage handling). The airport queuing system
has several characteristics which make it unfit for most traditional analysis:
1 time variation in arrival rate
2 dependence of service times on weather
3 passenger dependence.
 Passengers queue analysis in international airports terminals in Kerala 3

This study will focus on the M/M/C/∞ queuing system assume a Poisson arrival process.
This assumption is a very good approximation for arrival process in real systems that
meet the following rules:
1 the number of passengers in the system is very large
2 all passengers are independent, i.e., their decisions to use the system are independent
of other users.
This probability density distribution equation for a Poisson process describes the
probability of seeing n arrivals in a period from 0 to t.
(λ t ) n − λ t
Pn (t ) = e
n!
where
t is used to define the interval 0 to t
n is the total number of arrivals in the interval 0 to t
λ is the total average arrival rate in arrivals/sec.

1.3 Parameters in queuing models (multiple servers, multiple queues model)

n number of total passengers in the system (in queue plus in service)
c number of parallel servers
λ arrival rate (1/(average number of passengers arriving in each queue in a system))
μ serving rate (1/(average number of passengers being served at a server per hour))
cμ serving rate when c > 1 in a system
ρ system intensity or load, utilisation factor ( = l/(cμ)) (the expected factor of time the
server is busy that is, service capability being utilised on the average arriving
passengers).
Departure and arrival rate are state dependent and are in steady-state (equilibrium
between events) condition.

1.3.1 Objective of the study

In most of the airports, the major problem is waiting of passengers in the queue for more
deviation. This study analyse the queuing problem using alternative approach of the
multiserver method in the international departure of passengers in international airports in
Kerala. There are three international airports in Kerala namely Trivandrum, Kochi, and
Kozhikode. This study aims to study the existing queue pattern in these airports and to
suggest if any possible alternatives. Therefore to main objectives of this study are
1 to study the existing queue pattern of passengers in these airports
2 to identify the suitable queuing model for this state of art to fit a model
4 R. Aniyeri and R. Nadar

3 to develop an integrated queuing model that models passenger flow in the airport
terminal
4 to develop models with capability to determine resource utilisation at a high level of
detail.

2 Literature review

This literature review reveals various decision support techniques for the analysis of an
airport. The motivation for this study is to determine an appropriate model to solve the
queuing problem in international airports in Kerala. Numerous studies and work have
been done on the subject of airport. One of the studies carried out a simulation study of
the passenger check-in system at Ottawa International Airport. They fixed their attention
at the daily level and used simulation to determine: the minimal number of counters in
order to meet a service level for each separate flight in term of waiting times. Next, they
provided some integer programming formulations to minimise the total number of
counters and the total number of counter hours under the realistic constraint that counters
for one and the same flight should be adjacent. Another studied specifically the case of
the kiosks, or automated self-service check-in machines, and analysis has been done for
the improving of service quality of self-service kiosks, but also assisted the industry in
developing a SUSS (common-use self-service) standard that would enable airlines to
share kiosks. de Barros et al. (2007) analysed transfer passengers’ views on the quality of
services at the terminal building, using data collected at Bandaranaike International
Airport in Sri Lanka. In which regression analysis was used to identify the transfer
passenger facilities and services with the strongest effect on the overall perception of
level of service.
Another study about the check-in process of the airport terminal has been modelled
and examined. The goal of their study was to determine delays and solutions to improve
efficiency. A great number of security measures were adopted to avoid new terrorist
actions in airport terminals. Without proper security measures, people could consider the
air transportation system as unsafe and could refrain from travelling by aircraft. The state
of art overview highlights different research works in modelling airport terminal
operations. There are so many studies about Airport terminal using different method. One
study is about simulation model to analyse passenger and baggage flow in an airport
terminal. Another analyse the bottlenecks in the passengers flow and provide integral
solutions for supporting future airport developments. Another study is about baggage
screening strategies using artificial intelligence techniques. Again, another analyse the
problems related to the design and analysis of security screening and inspection system.
They introduces a new baggage-tracking system for improving airport security. In this
paper the authors propose a simulation model of the airport of Lamezia Terme (Calabria,
Italy) for investigating system performance under the effects of different scenarios
characterised by different resources allocation and availability. Many researchers have
faced issues concerning the optimisation of an airport terminal The Nagoya University
has conducted a simulation using the software Arena on departing flow passengers from
the International Kansai airport in Japan, in order to reduce the number of passengers,
because of long waiting times in peak periods and because of unavoidable delays, they
lose their flights. NAIA, Abuja-Nnamdi Azikiwe International used the death and birth
 Passengers queue analysis in international airports terminals in Kerala 5

rate approach to model to the waiting line at the airport and reported that more aircrafts
are needed on daily basis both on the domestic and international routes for improved
service. This study is a continuation of this past work aimed at developing a queuing
model using alternative approach of the multi-server method to facilitate the prediction,
processing of passengers at the airports in Kerala per operational period for
effectiveness. There are so many studies in the literature analysing the problem of
congestion in airport terminal. Guizzi et al. (2009) in their paper studied the analysis of
passenger flow in the terminal airport, from entrance to boarding. In particular this study
develops a simulation model based on the discrete event theory. A related study can be
found in Thiagaraj (2014) in his paper about the analytical approach and shows how to
build a simulation procedure. This study is to measure the performance of an airport
runway used only for arrivals, with different traffic mixes and operational variables.
Ashfaq (2006) in his paper studied analysis of the data at Birmingham Airport to provide
effective and useful information about the dwell-time that passengers have between
different points of their visit to the airport The method he uses a genetic paradigm which
is able to process that data using a compact and robust simulation model, so that the time
spent by the visitors to the airport can be extracted from the raw data produced by the
sensors. Barnhart (2003) in his paper studied several important areas of operations
research applications in the air transport industry. Mehri et al. (2008) in their paper
studied solving of waiting lines models in the airport using queuing theory model and
linear programming. The model about airport for passengers on a level with reservation is
the multiple-channel queuing model with Poisson arrival and exponential service times
(M/M/S). Total expected costs are studied, total costs is the sum of the cost of providing
service plus the cost of waiting time. Esteban and Pedro (2008) in their paper ‘Check in
process at Lisbon Airport’ used two methods are implemented, event-based simulations
and collaborative design.

2.1 Methodology
2.1.1 Data collection and sampling
There may be many possible extensions to this study. Therefore the data quality is crucial
and important for this study. A better and larger set of data should be obtained from
international airport in Kerala. Furthermore, an airport performance is crucial for
applying this technique. Thus, to determine more appropriate airport performance
variables like arrival of passengers and services in terminals using the different
definitions and derivations of the variables. By using the key variables selected from the
proposed methodologies, this paper reviews a queuing model for multiple servers. The
average queue length can be estimated simply from raw data collected through the
questionnaires, the number of passengers waiting in a queue in each minute. Then
compare this average with that of queuing model.
In order to compare the deviations,
• The data was collected directly from three airports in Kerala, India for a period of
one week by using the predetermined questionnaire.
• The observations were made for the number of passengers in a queue, their arrival-
time and departure-time without distracting the employees.
6 R. Aniyeri and R. Nadar

• The whole procedure of the service unit each day was observed and recorded using a
time-watch during the same time period for each day. In these airport terminals there
was no balk or renege.

2.1.2 Data analysis

The following assumptions were made for the analysis of the data.
1 The waiting line has two or more identical servers.
2 The arrivals follow a poison probability distribution with a mean arrivals rate of λ.
3 The service times follow an exponential probability distribution.
The mean service rate, μ is the same for each server.
4 The arrivals wait in a single line and then move to the first open server for service in
orderly manner.
5 The queue discipline is first-come-first-serve (FCFS).
6 As the traffic intensity ρ = λ/cμ gets closer to one, there is a very rapid increase in the
measures of congestion, Lq (expected number of passengers in the queue) and Wq
(expected time a passenger in the system) gets closer to one.
The multi server approach of modelling was adopted in this study to develop a
mathematical model to solve problem of queuing of air transport passengers at the
international airports in Kerala. The elementary probability theory to predict average
waiting time, average queue length, distribution of queue length, etc., on the basis of
‘arrival pattern of passengers’ to the resource, service pattern of the passengers’.

2.2 Components of queue models

Queues are not an unfamiliar phenomenon and to define it requires specification of the
characteristics which describes the system such as the arrival pattern, the service pattern,
the queue discipline and the queue capacity, Adedayo (2006) observed that there are
many queuing models that can be formulated. According to them it is essential that the
appropriate queuing model is used to analyse problems under study.
• The arrival pattern: This may be the arrival of an entity at a service point. This
process involves a degree of uncertainty concerning the exact arrival times and the
number of entities arriving. And to describe this process there are some important
attributes such as the sources of the arrivals, the size of each arrivals, the grouping of
such an arrival and the inter-arrival times.
• The service pattern: This may be any kind of service operation which processes the
arriving entities. The major features which must be specified are the number of
servers and the duration of the service.
• The queue discipline: This defines the rules of how the arrivals behave before service
occurs.
• The queue capacity: The queue capacity may be finite or infinite.
 Passengers queue analysis in international airports terminals in Kerala 7

2.3 Queuing model of international airports in Kerala

The system of operation at international airport terminals in Kerala can be model as a
queuing process. There are three international airports in Kerala I have to collect data
from three international airports in Kerala by questionnaire form for this research and
compare it. An attempt is made in this study to find solution to the queue of airport
terminal congestion in Kerala The problem can be modelled as a multi-server queue
problem with no system limit, arrival can be from a theoretically infinite source and the
service is on first-come-first-serve priority rule.

2.4 The arrival process of passengers

The passengers arrival process may be described in two ways
• by characterising the number of arrivals per units time (the arrival rate)
• by characterising the time between the successive arrivals (inter arrival time).
If λ is defined as the no: of arrivals per unit time then 1/λ will be the successive arrival. If
μ is the rate of service then 1/μ is the service time.
The average rate of passengers’ entering the queuing system as λ and the average rate
of serving passengers’ as μ a measure of traffic congestion for s server system is ρ = λ/s μ.
When ρ >1, the average no: of arrivals in to the system exceeds the maximum average
service rate of the system. To attain a steady state system ρ must less than one. When
ρ = 1 arrivals and service are deterministic therefore the average arrival rate and average
service rate guaranteed a steady state solution only if ρ < 1.
Our first goal is to study the property of a model of the arrival process to a system
and to compare its features to the Poisson process.
In determining whether the Poisson process is a reasonable model for arrivals in a
specific service system, it is useful to consider its three defining properties:
1 passengers arrive one at a time
2 the probability that a passengers arrives at any time is independent of when other
passengers’ arrive
3 the probability that a passengers arrives at a given time is independent of the time.
The assumption of a Poisson process will generally be a good one when the three
properties above are a reasonable description of the service system. However, it is
possible to perform more rigorous tests to determine if it is a good fit. The simplest tests
are based on the relationship of the standard deviation to the mean of the two
distributions involved in the Poisson process. Since the variance (square of the standard
deviation) of the Poisson distribution is equal to its mean, now examine the number of
arrivals in each fixed interval of time, and determine whether the ratio of the mean to the
variance is close to one. Alternatively, since the exponential distribution is characterised
by its standard deviation being equal to its mean, then look at the inter-arrival times and
compute the ratio of the standard deviation to the mean to see if it’s close to one.
The events occur successively in time, so that the intervals between successive events
are independently and identically distributed according to an exponential distribution.
8 R. Aniyeri and R. Nadar

Consider an arrival counting process {N(t) | t ≥ 0} denotes the total number of arrivals up
to time t, with N(t) = 0 and whish satisfies the following assumptions:
1 The probability that an arrival occurs between time t and time t + ∆t is equal to
λ ∆t + o(∆t) we write this Pr (arrival occurs between t and t + ∆t) where λ is a
constants independent of N(t), ∆t is an incremental element, and o(∆t) denotes a
quantity that becomes negligible when compared to ∆t as ∆t → 0 that is
o(Δt )
lim = 0.
Δt → 0 Δt

2 Pr (more than one arrival between t and t + ∆t) = o(∆t).

3 The number of arrivals in non overlapping intervals is statistically independent; i.e.,
the process has independent increments.
Now to calculate Pn(t), the probability of n arrivals in a time interval of length t, n being
an integer ≥ 0. For n ≥ 1 we have
Pn (t + Δt ) = Pr(n arrivals in t and none in Δt )
+ Pr(n − 1 arrivals in t and one in Δt )
+ Pr(n – 2 arrivals in t and two in Δt ) (1)
+ …………
+ Pr(no arrivals in t and n in Δt )

Using Assumptions 1 to 3 and equation (1) becomes

Pn (t + Δt ) = Pn (t ) [1 − λΔt – o(Δt ) ] + Pn −1 (t ) [ λΔt + o(Δt ) ] + o(Δt ) (2)

where the last term, o(∆t), represents the term Pr {n – j arrivals in t and j in (∆t);
2 ≤ j ≤ n}.
When n = 0, we have
P0 (t + Δt ) = P0 (t ) = [1 − λ Δt − o(Δt ) ] (3)

Rewriting equations (2) and (3) and combining all o(∆t) terms, we have
P0 (t + Δt ) − P0 (t ) = − λΔtP0 (t ) + o(Δt ) (4)

And
Pn (t + Δt ) − Pn (t ) = − λΔtPn (t ) + λΔtPn −1 (t ) + o(Δt ) (n ≥ 1) (5)

Dividing equations (4) and (5) by ∆t, and take the limit as ∆t → 0, and obtain the
differential – difference equation

⎡ P0 [t + Δt − P0 (t ) ] ⎤ o(Δt )
lim ⎢
Δt → 0 ⎣
⎥ = Δlim − λP0 (t ) + lim
Δt ⎦ t →0 Δt → 0 Δt

⎡ Pn ( t + Δt − Pn (t ) ) o(Δt ) ⎤
lim ⎢ = lim {− λPn (t ) + λPn −1 (t )} + lim ⎥
Δt → 0
⎣ Δ t Δ →0 Δ → 0 Δt
⎦
 Passengers queue analysis in international airports terminals in Kerala 9

This reduces to
dP0 (t )
= − λP0 (t ) (6)
dt
And
dPn (t )
= − λPn (t ) + λPn −1 (t ) (n ≥ 1) (7)
dt
Equation (6) clearly has the general solution P0(t) = Ce–λt, where the constants S is equal
to one, since P0(0) = 1 let n = 1 in equation (7) then
dP1 (t )
= − λP1 (t ) + λP0 (t )
dt
Or
dP1 (t )
+ λP1 (t ) = λP0 (t ) = λe − λtt
dt
The solution to this equation is given by
P1 (t ) = Ce− λt + λte− λt

Using the boundary condition Pn(0) = 0 for all n > 0 yields S – 0 and gives
P1 (t ) = λte− λt

Continuing sequentially to n = 2, 3, 4, …………….. in equation (7) and proceeding

similarly we get
( λt )2 − λt ( λt ) 2
P2 (t ) = e , P3 (t ) =
2 3!
and in general
( λt ) n − λt
Pn (t ) = e
n!
This is the formula for a Poisson distribution probability distribution with mean λt. Thus
the independently identically distributed random variable defined as the number of
arrivals to a queuing system by time t, this random variable has the Poisson distribution
with mean λt.
In order to show that if the arrival process is Poisson, an associated random variable
defined as the time between successive arrivals follows exponential distribution. Let T be
the random variable ‘time between successive arrivals’; then
Pr(T ≥ t ) = Pr(no arrivals in time t ) = P0 (t ) = e− λt

Therefore the cumulative distribution function of T can be written as

A(t ) = Pr (T ≤ t ) = 1 − e− λt
10 R. Aniyeri and R. Nadar

With corresponding density function

dA(t )
a(t ) = = λe− λt
dt
Thus T has the exponential distribution with mean 1/λ if the inter arrival time are
independent and have the same exponential distribution then the arrival follows the
Poisson distribution. So in airport terminal the arrival of passengers are independent.
Therefore, the arrival follows the Poisson distribution and service rate are exponentially
distributed.
There are s parallel service channels and a passengers can go to any of the free
counter for his service, where the service time is identical and have the same probability
density function s(t) with mean service rate μ per unit of time per busy server. Thus,
overall service rate when there are n units in the system is given by:
1 If n ≤ s, all the passengers’ may be served simultaneously and in such cases there
will be no queue. (s-n) servers may remain idle and then mean service rate is nμ,
for n = 0, 1,….., s – 1, s.
2 If n ≥ s, all the servers will remain busy, number of passengers’ waiting in the queue
will be (n-s) and then mean service rate is sμ, i.e., the mean service time μn is
given by
The mean service time μn is given time by
⎡ nμ if 0 ≤ n < s ⎤
µn = ⎢ ⎥
⎣ sμ if n ≥ s ⎦

2.5 Performance measures – measures of effectiveness

When analysis a queuing system, this study craving to find
• number of passengers in the system
• waiting time for a passenger
• the length of a busy or idle period
• the current work backload (in units of time).
These are called measures of effectiveness. They are all random variable.

2.6 Number of passengers

Let N denote the random variable that describes the number of passengers in the system
at steady state. The probability that at steady state the number of passengers’ present in
the system is n denoted by Pn
Pn = Prob[ N = n]

And average number in the system at steady state is

 Passengers queue analysis in international airports terminals in Kerala 11

∞
L = E[ N ] = ∑ np
n=0
n

Within the queuing system, passenger may be present in the queue waiting for the turn to
receive service or they may be receiving service. Let Nq be the random variables to
describe the number of passengers’ waiting in the queue and its mean is denoted by
Lq = E[Nq].

2.7 System time and queuing time

The time the passenger spends in the systems, from the instant of its arrival to the queue
to the instant of its departure from the server, is called the response time or sojourn time.
Let R denote the random variable that describes response time and its mean by E[R].
The response time composed of time the passenger spends waiting in the queue,
called the waiting time, plus the passengers spends receiving service, called the service
time. Let Wq be the random variables to describe the time the passengers spends waiting
in the queue and its mean is denoted by E[Wq].
This study analyses the application of queuing theory in international airports
passenger’s departures in the international airports in Kerala. In the airport terminals,
there are several servers in several queues. Therefore the apt queuing model is the
multiple servers in multiple queues of infinite capacity. Thus this study is a queuing
models of (M/M/S): (∞/FIFO).
This model deals with a queuing system having several servers arranged in parallel,
each of which has identically and independently distributed and service time is
distributed exponentially. The arrivals are assumed to follow Poisson distribution. The
system is assumed to have an infinite capacity and the passengers’ are served as first in
first out basis
Hence there are ‘S’ number of servers each with a mean service rate of μ and let the
mean arrival rate of passengers be λ.
The steady state probabilities for a Poisson queue system are given by
λ0 λ1 λ2 ………… λn −1
Pn = P0 where n ≥ 1 →1
μ1 μ2 μ3 ………… μn

Using the fact that

∞

∑P
n =0
n =1

We can write
∞
P0 + ∑P
n =1
n =1

Therefore
12 R. Aniyeri and R. Nadar

1
P0 = ∞
→2
λ0 λ1 λ2 λ3 ………… λn −1
1+ ∑
n =1
μ1 μ2 μ3 ………… μn −1

If there is a single server, the mean service rate μn = μ for all n. But for the given model,
there are s servers working independently of each other. When there are n passengers’ in
the system, the mean service rate, μn, san be calculated in two different situations:
1 If n < s, only n of the s servers will be busy and others will be idle. Hence, the mean
service rate will be nμ.
2 If n ≥ s, all the servers will be busy. Hence, the mean service rate will be sμ.
Thus, the following can be assumed.
1 the mean arrival time λn = λ for all n
2 the mean service time μn is given time by
⎡ nμ if 0 ≤ n < s ⎤
µn = ⎢ ⎥ →3
⎣ sμ if n ≥ s ⎦
3 the mean arrival rate is less than sμ, i.e., λ < sμ
If 0 ≤ n ≤ s, then substituting equation (3) in equation (1) we get
n
λ 1 ⎛λ⎞
Pn = P0 = ⎜ ⎟ P0 →4
1.μ.2 μ.3.μ …… nμ n!⎝ μ ⎠

If n ≥ s, then substituting equation (3) in equation (1) we get

λn
Pn =
(1.μ.2 μ.3 μ …… (.s − 1) μ )( sμ sμ …… (n − s + 1) times )
n
λn 1 ⎛λ⎞
Pn = = ⎜ ⎟ P0 →5
( s − 1)! μ ( sμ )n − s + 1 s ! s n − s ⎝ μ ⎠
s −1

To find the value of P0 we use the fast that

∞

∑P
n =0
n =1

i.e.,
⎡ s −1 1 ⎛ λ ⎞n ∞ 1 ⎛λ⎞ ⎤
n
⎢∑ ⎜ ⎟ + ∑ ⎜ ⎟ ⎥ P0 = 1
⎢⎣ n − 0 n ! ⎝ μ ⎠ n = s s ! s n − s ⎝ μ ⎠ ⎥⎦
 Passengers queue analysis in international airports terminals in Kerala 13

i.e.,
⎡ s −1 1 ⎛ λ ⎞n ∞ s s ⎛ λ ⎞n ⎤
⎢∑ ⎜ ⎟ + ∑ ⎜ ⎟ ⎥ P0 = 1
⎢⎣ n − 0 n ! ⎝ μ ⎠ n = s s ! ⎝ sμ ⎠ ⎥⎦
i.e.,
⎡ s −1 1 ⎛ λ ⎞ n s s ⎛ λ ⎞ s 1 ⎤
⎢∑ ⎜ ⎟ + ⎜ ⎟
⎢ n −0 n ! ⎝ μ ⎠ s ! ⎝ sμ ⎠ 1 − λ ⎥
⎥ P0 = 1
⎢⎣ μs ⎥⎦

λ
using the Assumption 3 that < 1 and the expansion is given by
sμ

∞ n
⎛ λ ⎞ ⎛ λ ⎞
s
⎡ λ ⎛ λ ⎞
2
⎤ ⎛ λ ⎞s 1
∑ ⎜ ⎟ =⎜ ⎟
sμ ⎠ ⎝ sμ ⎠
⎢1 + + ⎜ ⎟ + ……⎥ = ⎜ ⎟
μs ⎝ μs ⎠ μs λ
n= s ⎝ ⎣⎢ ⎦⎥ ⎝ ⎠ 1 −
μs

Thus it follows that

⎡ s −1 1 ⎛ λ ⎞ n 1 ⎛λ⎞ ⎤
s
⎢∑ ⎜ ⎟
⎢ n =0 n ! ⎝ μ ⎠
+
⎛
⎜ ⎟ ⎥ P0 = 1
λ ⎞⎝ μ⎠ ⎥
⎢ s !⎜ 1 − ⎟ ⎥
⎣ ⎝ μs ⎠ ⎦
Hence
1
P0 = n s
→6
s −1
1 ⎛λ⎞ 1 ⎛λ⎞
∑ ⎜ ⎟ +
n! ⎝ μ ⎠ ⎛
⎜ ⎟
λ ⎞ μ
n=0 s !⎜ 1 − ⎟ ⎝ ⎠
⎝ sμ ⎠

3 Characteristics of the model

Expected number of passengers’ in the queue or queue length (Lq). This model has a total
of s servers. The expected queue length is given by
∞ ∞
Lq = E ( N q ) = E ( N – s ) = ∑
n=s
(n − s ) Pn = ∑ xP
x =0
x+ s putting x = n − s

Thus we have
∞ s+ x s ∞ x
x ⎛λ⎞ 1⎛λ⎞ ⎛ λ ⎞
Lq = ∑
x=0
⎜ ⎟
s !s x ⎝ μ ⎠
P0 = ⎜ ⎟
s!⎝ μ ⎠
P0∑ x⎜ ⎟
x =1 ⎝
sμ ⎠
14 R. Aniyeri and R. Nadar

it follows that
s
1⎛λ⎞ λ 1
Lq = ⎜ ⎟ P0 2
s!⎝ μ ⎠ μs ⎛ λ ⎞
⎜ 1 − ⎟
⎝ μs ⎠
Using the fact that
∞ x ∞ x −1 −2
⎛ λ ⎞ λ ⎛ λ ⎞ λ ⎡ λ⎤
∑ x⎜ ⎟ =
x =1 ⎝
sμ ⎠ μs ∑ x⎜ ⎟
x =1 ⎝
μs ⎠
= ⎢1− ⎥
μs ⎣ sμ ⎦

Hence we deduce that

s +1
⎛λ⎞
⎜ ⎟
Lq =
1 ⎝μ⎠ P
2 0
ss ! ⎛ λ ⎞
⎜1 − ⎟
⎝ μs ⎠

3.1 Average or expected number of passengers’ in the system (Ls)

By Little’s formula we have
s +1
⎛λ⎞
⎜ ⎟
Ls = Lq +
λ
=
1 ⎝ μ⎠
P +
λ
2 0
μ ss ! ⎛ λ ⎞ μ
⎜1 − ⎟
⎝ μs ⎠
where
∞
Ls = E[ N ] = ∑ nP
n =0
n

3.2 Average waiting time of a passenger in the system (Ws) By Little’s formula
s +1
⎛λ⎞
⎜ ⎟
Ws =
Ls 1 1
= ∗ P + ⎝ μ⎠
1
2 0
λ λ ss ! ⎛ λ ⎞ μ
⎜1 − ⎟
⎝ μs ⎠
i.e.,
 Passengers queue analysis in international airports terminals in Kerala 15

s
⎛λ⎞
⎜ ⎟
Ws =
1 1 1
+ ⎝ μ⎠
P
2 0
μ μ ss ! ⎛ λ ⎞
⎜1 − ⎟
⎝ μs ⎠

3.3 The average waiting time of a passenger in the queue (Wq)

By Little’s formula
s +1
⎛λ⎞
⎜ ⎟
Wq =
Lq 1 1
= ⎝ μ⎠ P
2 0
λ λ ss ! ⎛ λ ⎞
⎜1 − ⎟
⎝ μs ⎠
s
⎛λ⎞
⎜ ⎟
Wq =
1 1 ⎝ μ⎠ P
2 0
μ ss ! ⎛ λ ⎞
⎜1 − ⎟
⎝ μs ⎠

3.4 The probability that an arrival has to wait for service

The arrival has to wait for service if and only if Ts > 0 where Ts denote the waiting time
of a passenger in the system, i.e., if and only if there are s or more passengers’ in the
system. Thus the required probability is equal to
∞
P (Ts > 0 ) = P ( N ≥ s ) = ∑P
n=s
n

∞ n ∞ n n−s
1 ⎛λ⎞ 1⎛λ⎞ ⎛ λ ⎞
= ∑
n=s
n−s ⎜
s !s ⎝ μ ⎠
⎟ P0 = ⎜ ⎟
s!⎝ μ ⎠
P0 ⎜ ⎟
n=s ⎝
∑
μs ⎠

i.e.,
s
⎛λ⎞
⎜ ⎟ P0
μ
P (Ts > 0 ) = P ( N ≥ s ) = ⎝ ⎠
⎛ λ ⎞
s !⎜1 − ⎟
⎝ sμ ⎠
16 R. Aniyeri and R. Nadar

3.5 The probability that an arrival enters the service without waiting
s
⎛λ⎞
⎜ ⎟ P0
μ
1 − P (Ts > 0 ) = 1 − ⎝ ⎠
⎛ λ ⎞
s !⎜1 − ⎟
⎝ sμ ⎠

3.6 The mean waiting time in the queue for those who need to wait

E [Tq ]
E [Tq | Ts > 0] =
P [Ts > 0]
Wq
=
P [Ts > 0]

Substituting the values of Wq and P [Ts > 0] in the above equation we get
s
⎛λ⎞ ⎛ λ ⎞
⎜ ⎟ s !⎜1 − ⎟
E [Tq | Ts > 0] =
1 1 ⎝ μ⎠ P× ⎝ s ⎠
μs
2 0
μ ss ! ⎛ λ ⎞ ⎛λ⎞
⎜1 − ⎟ ⎜ ⎟ P0
⎝ μs ⎠ ⎝μ⎠
Simplifying we get
1
E [Tq | Ts > 0] =
⎛ λ ⎞
μs ⎜1 − ⎟
⎝ μs ⎠
1
=
μs − λ

3.7 The average or expected number of passengers’ in non empty queues (Ln)
If N denote s the number of passengers’ in the system and Nq denote the number of
passengers’ in the queue, then Ln is the conditional expectation defined by
Ln = E [ N q | N q ≥ 1]
E [ Nq ]
=
P[ N ≥ s ]
Lq
=
P[ N ≥ s ]
 Passengers queue analysis in international airports terminals in Kerala 17

Since there are s servers, substituting the values of Lq and P [N ≥ s] we get

s +1
⎛λ⎞ ⎛ λ ⎞ ⎛ λ ⎞
⎜ ⎟ s !⎜1 − ⎟ ⎜ ⎟
Ln =
1 ⎝ μ⎠ μs
P× ⎝ s ⎠= ⎝ ⎠
sμ
2 0 λ
ss ! ⎛ λ ⎞ ⎛ λ ⎞ 1−
⎜1 − ⎟ ⎜ ⎟ P0 sμ
⎝ μs ⎠ ⎝ sμ ⎠
Now applying the above multiserver queuing formula from the data collected from the
three airports in Kerala (see the Appendix).

4 Data collection at international airports in Kerala

• Data collection at Kozhikode: The path related to passengers departing from

Terminal 1 consists of two floors: on the ground floor we have the check-in desks,
and on the first floor, assessable by escalators, the security control facilities
immigration, security and customs and some shops and food services are located. On
the other side of the ground floor besides the path related to arriving passengers there
are the ticket office, the two currency exchange offices, several car rentals, and so
forth. Simulation studies require an exact description of processes and representative
data. Therefore correct data is essential to get valid and valuable results about
bottlenecks and to define relevant scenarios.
• Data collection at Kochi: The country’s first green field airport in the public private
partnership model has integrated some unique and standout features in to its terminal
design that marks the biggest expansion program since the airport turned operational
in 1999. The present international terminal with a built-up space of 4.75 lakh sq.ft
has a peak hour capacity to handle only 1,200 passengers each on the arrival and
departure sides. The terminal’s elevation has been modelled along the State’s
traditional and aesthetically-rich temple architecture. It would be a multi-level
terminal complex with the ground level to feature arrivals, and departures in the first
level.
• Data collection at Trivandrum: The path related to passengers departing from
Terminal 1 consists of two floors: on the ground floor we have the check-in desks,
and on the first floor, assessable by escalators, the security control facilities
immigration, security and customs and some shops and food services are located. On
the other side of the ground floor besides the path related to arriving passengers there
are the ticket office, the two currency exchange offices, several car rentals, and so
forth. simulation studies require an exact description of processes and representative
data. Therefore correct data is essential to get valid and valuable results about
bottlenecks and to define relevant scenarios.
18 R. Aniyeri and R. Nadar

5 Data analysis and computation at international airports in Kerala

The most problematic phase of this study has been that on data acquisition. Sampling
took place in the passenger terminal Karipur International airport. After a thorough
inspection airport it same out that passenger traffic peaks during Wednesday. Therefore
the focus data collection phase was on three days: Monday, Wednesday and Friday. Data
was grouped in the spreadsheets that show some information such as basis information
flight (the airline and destination), time of departure, time of opening and closing its
check-in and finally the flight code. This information is important because from the flight
code it is possible to know the type of the aircraft used for that flight and then the total
number of aircraft available seats. In order to carry out a correct analysis of collected data
it was necessary to make an appropriate data stratification for highlighting some key
aspects. By observation the following data was collect at exactly 7.00 am to 1.00 pm on
daily basis. Table shows the arrival, inter-arrival, service for a certain group of random
passengers that used the facilities in airport terminal as at the time of observation and
data collection.
Sampling took place in the passenger terminal Kochi International Airport. After a
thorough inspection airport it same out that passenger traffic peaks during midnight. Data
was grouped in the spreadsheets that show some information such as basis information
flight (the airline and destination), time of departure, time of opening and closing its
check-in and finally the flight code. This information is important because from the flight
code it is possible to know the type of the aircraft used for that flight and then the total
number of aircraft available seats. In order to carry out a correct analysis of collected data
it was necessary to make an appropriate data stratification for highlighting some key
aspects. By observation the following data was collect at exactly 7.00 am to 5.00 pm on
daily basis. Table shows the arrival, inter-arrival, service for a certain group of random
Passengers that used the facilities in airport terminal as at the time of observation and
data collection
Sampling took place in the passenger terminal Trivandrum International Airport.
After a thorough inspection airport it same out that passenger traffic peaks during
Midnight. Data was grouped in the spreadsheets that show some information such as
basis information flight (the airline and destination), time of departure, time of opening
and closing its check-in and finally the flight code. This information is important because
from the flight code it is possible to know the type of the aircraft used for that flight and
then the total number of aircraft available seats. In order to carry out a correct analysis of
collected data it was necessary to make an appropriate data stratification for highlighting
some key aspects. By observation the following data was collect at exactly 12.00 am to
2.00 am on daily basis. Table shows the arrival, inter-arrival, service for a certain group
of random passengers that used the facilities in airport terminal as at the time of
observation and data collection.

5.1 Chart diagram of international airports in Kerala

In both the charts shows that in peak hours the passengers in the queue and waiting
passengers in the queue is very high. This shows that the efficiency of servers in Dubai
and Sharjah airlines has to increase.
 Passengers queue analysis in international airports terminals in Kerala 19

Figure 1 Chart of Karipur Airport, (a) average number of passengers’ in the queue (Lq)
(b) average waiting time of passengers in the queue (see online version for colours)

(a)

(b)
20 R. Aniyeri and R. Nadar

Figure 2 Chart of Kochi Airport, (a) average number of passengers’ in the queue (Lq)
(b) average waiting time of passengers in the queue (see online version for colours)

(a)

(b)

Cochin International Airport also known as Nedumbassery Airport and CIAL, is the
largest and busiest airport in Kerala, India located at Nedumbassery,. In India it is the
fourth busiest in terms of international passenger traffic. Cochin International Airport
(CIAL) is the first Greenfield airport setup in the public private partnership (PPP) model
in civil aviation infrastructure sector in India. From the above graph shows that in every
airlines average number of passengers in the queue and average number of passengers
waiting in the queue is very high.
 Passengers queue analysis in international airports terminals in Kerala 21

Figure 3 Chart of Trivandrum Airport, (a) this is the chart of average number of passengers’ in
the queue (Lq) (b) this is the chart of average waiting time of passengers in the queue
(see online version for colours)

(a)

(b)

From the above charts it shows that there will be a big queue in the airport in peak hours.
Now I switch on to scrutinising the above data’s used in multiserver queuing theory, by
applying the analysis of variance.
22 R. Aniyeri and R. Nadar

6 Analysis of variance

Now this study moves on to the area of Analysis of variance using two way
classifications. Analysis of variance, otherwise known as ANOVA, is a collection of
statistical models used to analyse the differences among groups and their associated
procedures (such as variation among and between groups), developed by statistician and
evolutionary biologist Ronald Fisher test.
Two-way analysis of variance performs an analysis of variance for testing the equality
of populations means when classification of treatments is by two variables or factors. In
two-way ANOVA, the data must be balanced (all cells must have the same number of
observations) and factors must be fixed.Two-way ANOVA model works in balanced
case only. That is we need to have the same number of observations (replications) for
each factor levels combinations.

7 Notations and computation

The data for two-factor ANOVA can be displayed in a two-way table

• no. of levels rows of factor = r
• no. of levels of columns factor = c
• total no. of observations = rc
• observation in (ij)th cell of the table = xij
• ith level of row factor and i = 12……. r
• jth level of column factor j = 12………
Sum of c observation in the ith row

= TRi = ∑x i
ij

Sum of r observation in the jth column

= TCi = ∑x i
ij

Sum of all rc observations

T= ∑∑ x = ∑ T = ∑ T
j i
ij
i
Ri
j
Ci

This lead to following computation

Total sum of squares
T2
∑∑ X
2
SST = IJ −
I J
rc
 Passengers queue analysis in international airports terminals in Kerala 23

Between rows sum of squares

TRi 2 T 2
SS R = ∑
i
c
−
rc

Between columns sum of squares

TCi 2 T 2
SSC = ∑
i
r
−
rc

Error sum of squares

SS E = SST – SS R – SSC .
Table 1 ANOVA table and hypothesis test

Source of Sum of Degrees of Mean

F ratio
variation squares freedom squares
Between the SSR r–1 MSR MS R
rows MS E
Between the SSC c–1 MSC MS R
columns MS E
Error SSE (r – 1) (c – 1) MSE
Total SST rc – 1

7.1 ANOVA table and hypothesis test

Conditions
1 three sum of squares are SSR, SSC, SSE are independently distributed
2 degree of freedom r – 1 + c – 1+ (r – 1) (c – 1) = rc – 1.
Using the F ratio tests for significant row effects and for significant column effects can be
under taken.
H0: no effect due to row factor H0: no effect due to column factor
H1: effect due to row factor H1: effect due to column factor
Critical region Critical region
F > Fαr–1,(r–1)(c–1) F > Fαc–1,(r–1)(c–1)
Test statistic Test statistic

7.1.1 Data analysis by two ways classification in Karipur Airport

The null hypothesis is that there is no significant difference between the terminals as well
as airlines. The level of significance is 5%.Here the table value of F(3,12) = 3.49 and
F(4,12) = 3.26. Here FT < Fα (table value) and FA < Fα (table value) means accept the
hypothesis. Therefore concluded that lines in various points seems to be different. This
may be due to the service time. If the service time decreases the length of the waiting line
24 R. Aniyeri and R. Nadar

or queue length can be minimised. These are all depending up on the efficiency of the
service persons and infrastructure of air service.
Table 2 Two-way classifications in Lq of Karipur Airport

F ratio
Source of Sum of Degree of Mean sum
variation squares freedom of squares F table
F calculated
0.05
Between the SST = 39.25511 c–1=3 MST = SSC/D.F FT = 1.324 3.49
terminals = 13.085
Between the SSA = 75.7178 r–1=4 MSA = SSR/D.F FA = 1.916 3.26
airlines = 18.9294
Error SSE = 118.5731 12 MSE = SSE/D.F
= 9.88109
Total 233.54 19

The null hypothesis is that there is no significant difference between the terminals as well
as airlines. The level of significance is 5%. Here the table value of F(3,12) = 3.49 and
F(4,12) = 3.26. Here FT < Fα (table value) accept the hypothesis and FA > Fα (table value)
reject the hypothesis. Therefore we concluded that lines in various points seem to be
different. This may be due to the service time. If the service time decreases the length of
the waiting line or queue length can be minimised. These are all depending up on the
efficiency of the service persons and infrastructure of air service.
Table 3 Two-way classification in Wq of Karipur Airport

Source of Sum of Degree of Mean sum F ratio

variation squares freedom of squares F calculated F table
Between the SST = 398.5486 c–1=3 MST = SSC/D.F FT = 1.015 3.49
terminals = 132.8495
Between the SSA = 1,948.7612 r–1=4 MSA = SSR/D.F FA = 3.612 3.26
airlines = 487.194
Error SSE = 1,618.6976 12 MSE = SSE/D.F
= 134.8925
Total 3,952.8965 19

7.1.2 Data analysis by two-way classification in Kochi Airport

The null hypothesis is that there is no significant difference between the terminals as well
as airlines. The level of significance is 5%. Here the table value of F(3,12) = 3.49 and
F(4,12) = 3.26. Here FT < Fα (table value) and FA < Fα (table value) means accept the
hypothesis. Therefore, we concluded that lines in various points seem to be different.
This may be due to the service time. If the service time decreases, the length of the
waiting line or queue length can be minimised. These are all depending up on the
efficiency of the service persons and infrastructure of air service.
 Passengers queue analysis in international airports terminals in Kerala 25

Table 4 Two-way classification in Lq of Kochi Airport

F ratio
Source of Sum of Degree of Mean sum
variation squares freedom of squares F table
F calculated
0.05
Between the SST = 8.829133 c–1=3 MST = SSC/D.F FT = 1.21 3.49
terminals = 2.943044
Between the SSA = 21.474199 r–1=4 MSA = SSR/D.F FA = 1.5068 3.26
airlines = 5.36854995
Error SSE = 42.753625 12 MSE = SSE/D.F
= 3.56280208
Total 73.056958 19

The null hypothesis is that there is no significant difference between the terminals as well
as airlines. The level of significance is 5%. Here the table value of F(3,12) = 3.49 and
F(4,12) = 3.26. Here FT < Fα (table value) and FA < Fα (table value) means accept the
hypothesis. Therefore, we concluded that lines in various points seem to be different.
This may be due to the service time. If the service time decreases, the length of the
waiting line or queue length can be minimised. These are all depending up on the
efficiency of the service persons and infrastructure of air service.
Table 5 Two-way classification in Wq of Kochi Airport

F ratio
Source of Sum of Degree of Mean sum
variation squares freedom of squares F table
F calculated
0.05
Between the SST = 278.97521 c–1=3 MST = SSC/D.F FT = 1.7616 3.49
terminals = 92.99173
Between the SSA = 1,911.17857 r–1=4 MSA = SSR/D.F FA = 2.9167 3.26
airlines = 477.79565
Error SSE = 1,965.79543 12 MSE = SSE/D.F
= 163.81628
Total 4,155.949212 19

7.1.3 Data analysis by two-way classification in Trivandrum Airport

The null hypothesis is that there is no significant difference between the Terminals as
well as airlines. The level of significance is 5%. Here the table value of F(3,12) = 3.49
and F(4,12) = 3.26. Here FT < Fα (table value) accept the hypothesis and FA > Fα (table
value) reject the hypothesis. Therefore, we concluded that lines in various points seem to
be different. This may be due to the service time. If the service time decreases the length
of the waiting line or queue length can be minimised. These are all depending up on the
efficiency of the service persons and infrastructure of air service.
26 R. Aniyeri and R. Nadar

Table 6 Two-way classification in Lq of Trivandrum Airport

F ratio
Source of Sum of Degree of Mean sum
variation squares freedom of squares F table
F calculated
0.05
Between the SST = 46.12254 c–1=3 MST = SSC/D.F FT = 1.6466 3.49
terminals = 15.37418
Between the SSA = 335.7423 r–1=4 MSA = SSR/D.F FA = 8.9895 3.26
airlines = 83.93558
Error SSE = 112.0447759 12 MSE = SSE/D.F
= 9.33706
Total 493.909617 19

The null hypothesis is that there is no significant difference between the Terminals as
well as airlines. The level of significance is 5%. Here the table value of F(3,12) = 3.49
and F(4,12) = 3.26. Here FT > Fα (table value) reject the hypothesis, and FA < Fα (table
value) accept the hypothesis. Therefore, we concluded that lines in various points seem to
be different. This may be due to the service time. If the service time decreases, the length
of the waiting line or queue length can be minimised. These are all depending up on the
efficiency of the service persons and infrastructure of air service.
Table 7 Two-way classification in Wq of Trivandrum Airport

F ratio
Source of Sum of Degree of Mean sum
variation squares freedom of squares F table
F calculated
0.05
Between the SST = 1,437.88403 c–1=3 MST = SSC/D.F FT = 3.9185 3.49
terminals = 479.1119
Between the SSA = 543.306329 r–1=4 MSA = SSR/D.F FA = 1.1108 3.26
airlines = 135.8266
Error SSE = 1,467.241942 12 MSE = SSE/D.F
= 122.2702
Total 3,447.8840314 19

8 Results and discussion

A number of studies have performed passenger queue simulations in airport terminals for
the purpose of analysing current and future levels of service. Traditionally these studies
have focused on the mandatory processing facilities such as check-in, security,
immigration, airside boarding lounge and the boarding gates themselves through the
application of queueing theory.
The queue problem at international airports in Kerala was modelled as a multi-server
queuing problem. Theoretically, arrival is from infinite source and the service pattern is
on FCFS. One week data of departures of passengers at international airports in Kerala is
considered. See Tables A1 to A3, Adebayo (2008) stated that the closer the traffic
intensity it to zero the more efficient the operations of the service facilities. According to
Satty (1961), if λ > 1, the number of passengers would be infinite. In the International
 Passengers queue analysis in international airports terminals in Kerala 27

Airport Karipur in peak hours the average no of customers in the queue is more than
20 per minute and in the non-peak hours is more than ten per minute. In the International
Airport Karipur in peak hours the average no of customers waiting in the queue is more
than 45 per minute and in the non-peak hours is more than 28 per minute. In the
International Airport Kochi in peak hours the average no of customers in the queue is
more than 25 per minute and in the non-peak hours is more than 15 per minute. In the
International Airport Kochi in peak hours the average no of customers waiting in the
queue is more than 1 1/2 hour and in the non-peak hours is more than one hour. In the
International Airport Trivandrumin peak hours the average no of customers in the queue
is more than 25 per minute and in the non-peak hours is more than ten per minute. In the
International Airport Trivandrum in peak hours the average no of customers waiting in
the queue is more than 1 and 1/2 hours and in the non-peak hours is more than 50-minute
(Tables A1 to A3). This shows that that during weekdays prime hours there is heavy
passengers in terminals.

9 Conclusions

In the case of Dubai Airlines, passengers take more time for baggage than other activities
when compared to other air services. This may be due to their liberalism in baggage
restriction.
This study has been done by observing the passengers arrival time, waiting time in
the queue, different behaviour of passengers in the queue like jockeying and service time
in airport terminals in Kerala. This study was made the information for one week
duration in all the three airport during weekdays peak and non-peak hours. Generally,
arrivals do not occur at fixed regular intervals of times but tend to be clustered for a
duration of a week. The Poisson distribution involves the probability of occurrence of an
arrival are random and independent of all other operating conditions. The inter arrival
rate (i.e., the number of arrivals per unit of time) λ is calculated by considering arrival
time of the customers to that of the number of customers. Service time is the time
required for completion of a service, i.e., it is the time interval between beginning of a
service from airport and its completion. In this study the researcher has calculated mean
service time μ of passengers by considering different service time of passengers. Based
upon the tabulation and taking one day as a standard, the researcher inferred that during
prime hours there is heavy passengers in terminals, which implies that the utilisation
factor is 1. With the increasing number of customers’ coming to airport terminal, there
must be trained employees serving at each service unit. Increasing more than sufficient
number of servers may not be the solution to increase the efficiency of the server by each
service unit.

References
Ashfaq, W. (2006) ‘Analysis of passenger movement at Birmingham International Airport’,
Proceedings of IEEE World Conference on Computational Intelligence (IEEE Congress on
Evolutionary Computation), pp.1339–1345.
Barnhart, C. (2003) ‘Applications of operations research in the air transport industry transportation
science’, INFORMS, November, Vol. 37, No. 4, pp.368–391.
28 R. Aniyeri and R. Nadar

Esteban, D. and Pedro, J. (2008) ‘Check-in process at Lisbon Airport’, PhD thesis, Event-based
Simulations and Collaborative Design.
Guizzi, G., Murino, T. and Romano, E. (2009) ‘A discrete event simulation to model passenger
flow in the airport terminal’, MMACTEE'09 Proceedings of the 11th WSEAS International
Conference on Mathematical Methods and Computational Techniques in Electrical
Engineering, pp.427–434.
Mehri, H., Djemel, T. and Kammoun, H. (2008) Solving of Waiting Lines Models in the Airport
Using Queuing Theory Model and Linear Programming the Practice Case, 25 pp, 11 March,
2006, HAL ID: hal-00263072.I.M.H.B, hal-00263072v2.
Thiagaraj, H.B. (2014) ‘A queueing model for airport capacity and delay analysis’, Applied
Mathematical Sciences, Vol. 8, No. 72, pp.3561–3575.
Van der Sluis, E. and Van Dijk, N.M. (2006) ‘Check-in computation and optimization by
simulation and IP in combination’, European Journal of Operational Research, June,
Vol. 171, No. 3, pp.1152–1168.

Appendix

Table A1 Data analysis of Karipur International Airport

Dubai
Baggage Immigration Customs Security
Po 0.01475 0.03774 0.16667 0.17266
Lq 5.16465 1.52847 4.16675 2.17074
Wq 10.3293 4.58541 16.66667 6.51222
Saudi
Baggage Immigration Customs Security
Po 0.01475 0.0008029 0.06667 0.044944
Lq 5.1647 2.5125 5.71695 3.51125
Wq 10.3294 2.513 22.87 17.56
Dubai
Baggage Immigration Customs Security
Po 0.0598 0.00812 0.09091 0.0562
Lq 3.027 12.2020 3.788 2.5
Wq 9.081 12.2020 11.364 12.5
Saudi
Baggage Immigration Customs Security
Po 0.9091 0.009140 0.14285 0.0747
Lq 3.788 1.265 1.928 1.7033
Wq 11.364 2.53 7.712 6.813
Sharjah
Baggage Immigration Customs Security
Po 0.053697 0.00585 0.052632 0.0641
Lq 14.4 7.13451 7.673454 2.14
Wq 57.6 35.40699 38.36727 6.4298
 Passengers queue analysis in international airports terminals in Kerala 29

Interpretation of results for queuing model 1

The interpretation of the airport terminal analysis of Karipur International Airport from
above table is that the probability for servers to be busy in Dubai flight-baggage
screening is 0.1475, i.e., 14.75%. The average number of passengers waiting in a queue is
Lq = 5.16465 passengers per minutes. The waiting time in a queue per server is
Wq = 10.33 minutes.
From the flight code, it is possible to know the type of the aircraft used for that flight
and then the total number of aircraft available seats. In order to carry out a correct
analysis of collected data it was necessary to make an appropriate data stratification for
highlighting some key aspects.
Table A2 Data analysis of Kochi International Airport

Colombo
Baggage Immigration Customs Security
Po 0.04889 0.5769 0.09677 0.05101
Lq 5.133 6.8560 4.03208 5.0756
Wq 40.24 11.112 12.09624 36.56
Abudhabi
Baggage Immigration Customs Security
Po 0.1091 0.00895 0.04348 0.05394
Lq 7.2758 6.87446 9.6453 4.99051
Wq 43.5211 42.6100 57.8700 32.66
Dubai
Baggage Immigration Customs Security
Po 0.04708 0.02548 0.07527 0.03450
Lq 5.194 5.970 9.29047 5.61
Wq 41.552 42.310 65.033 30.05
Riyadh
Baggage Immigration Customs Security
Po 0.01473 0.10470 0.16271 0.0756
Lq 6.498 6.499 2.20336 3.3356
Wq 29.24 24.177 17.62688 13.3424
Dubai
Baggage Immigration Customs Security
Po 0.03439 0.1652 0.14286 0.05125
Lq 5.6184 6.39 1.9286 4.562
Wq 32.1602 45.287 19.286 22.81
30 R. Aniyeri and R. Nadar

Table A3 Data analysis of Trivandrum International Airport

Saudi
Baggage Immigration Customs Security
Po 0.0345 0.0000928 0.1659 0.125
Lq 2.5557 4.60767 3.455 2.382
Wq 10.288 5.148 24.185 21.438
Dubai
Baggage Immigration Customs Security
Po 0.0556 0.000210 0.0526 0.04773
Lq 21.476 18.168 7.6736 8.634
Wq 106.6 18.168 38.36 42.220
Doha
Baggage Immigration Customs Security
Po 0.148 0.0000374 0.090909 0.10345
Lq 5.165 6.8346 3.7878 3.156
Wq 10.33 6.12 34.0902 25.258
Abudhabi
Baggage Immigration Customs Security
Po 0.0471 0.000113 0.07359 0.07692
Lq 5.13 7.18289 5.03141 4.747
Wq 40.24 7.18289 11.0692 33.23
Dubai
Baggage Immigration Customs Security
Po 0.12987 0.000036 0.013243 0.1111
Lq 2.2165 2.247 3.580 2.84
Wq 6.6495 2.247 7.55434 28.4