Download as pdf or txt
Download as pdf or txt
You are on page 1of 17

Journal of Air Transport Management 59 (2017) 83e99

Contents lists available at ScienceDirect

Journal of Air Transport Management


journal homepage: www.elsevier.com/locate/jairtraman

Airline new route selection based on interval type-2 fuzzy MCDM:


A case study of new route between Turkey- North American region
destinations
Muhammet Deveci a, *, Nihan Çetin Demirel a, Emine Ahmetog
lu a
a
University of Yildiz Technical, 34349 Yildiz, Istanbul, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: Conditions of global competition in the aviation sector force airlines to be internationally-minded instead
Received 28 August 2016 of competing only in their domestic markets. Thus, airlines are interested in potential international
Received in revised form markets instead of domestic markets. This study focuses on the process of new route selection, which has
12 October 2016
a high impact on the revenue and passenger numbers of airlines. An airline company in Turkey plans to
Accepted 26 November 2016
launch a new route at an airport in the North American region. Any suboptimal decision has a huge effect
on the outcomes of airlines in the market. Therefore, making an optimal decision, which is compatible
with the airline company's goal, is highly important. In this study, a decision analysis is applied in terms
Keywords:
New route analysis
of the selection of a new route from five different destinations by using an interval type-2 fuzzy TOPSIS
Interval type-2 fuzzy sets method. A multi-criteria decision-making (MCDM) method is used in order to identify the aspects of the
TOPSIS new route's feasibility. Finally, a real case study is shown to verify the proposed method and to
Airline industry demonstrate its practicality and feasibility. The results show that the MCDM approach is a useful tool for
Airport decision-makers in terms of selecting potential airports for route analysis.
Multi-criteria decision making (MCDM) © 2016 Elsevier Ltd. All rights reserved.

1. Introduction started to join the sector globally (THY, 2009). The aviation in-
dustry, which plays a key role in the development and competi-
The sector of aviation has been on the rise, especially after the tiveness of countries, has developed around the world in parallel
second half of the 1980s. During this period, it has been observed with economic growth, globalisation, and liberalization. Since
that some airline companies have started a modernization and 2002, world air traffic has increased by 49% in terms of revenue
standardization process in an effort to improve and elevate service passenger miles. Turkey has been among the four fastest-growing
standards. The emphasis has been on international flights, which countries in the world in terms of civil aviation for the last eight
are more economically advantageous than domestic ones. At the years (THY, 2011).
same time, a substantial increase in the number of airline com- Managing and regulating Turkish airports and controlling
panies, their fleet capacities, and their share in the market has been Turkish airspace are carried out by DHMI, a state-owned enterprise
observed (DPT, 2001). operating under Ministry of Transport, Maritime Affairs and Com-
In Turkey today, air travel has drawn more attention than other munications. Inviting tenders for construction of new terminals,
modes of traveling. The aviation sector has experienced constant runways, and airports is among the regulations of this enterprise.
progress and change since the early years, when air travel started in DHMI invited tenders for the third airport of Istanbul in May 2013.
earnest. Since 2003, the progress and change have gained mo- It will be one of the greatest international airports in the world
mentum, and an increase in the number of private airline com- when the fourth and final phase is finished by 2028. It will be also
panies has been noted. In recent years, low-cost airlines have one of the busiest airports in the world with six runways, flights to
nearly 350 destinations and an eventual annual capacity of 200
million passengers. When first phase is completed in 2018, Istan-
* Corresponding author.
bul's new airport will have the world's largest terminal under one
E-mail addresses: muhammetdeveci@gmail.com, mdeveci@yildiz.edu.tr roof, including a gross floor area of nearly 11 million square feet
(M. Deveci), nihan@yildiz.edu.tr (N.Ç. Demirel), eahmetoglu10@gmail.com with the ability to serve 90 million passengers annually.
(E. Ahmetoglu).

http://dx.doi.org/10.1016/j.jairtraman.2016.11.013
0969-6997/© 2016 Elsevier Ltd. All rights reserved.
84 M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99

Constructed on the Black Sea coast on the European side of Turkey, 2 sets. During this process, five different airports have been eval-
the airport will be 22 miles outside Istanbul. Istanbul's new airport uated according to eleven criteria by five decision-makers and have
will become Turkey's main airport and a hub for connecting flights been sorted according to the calculated indices. This study shows
between Europe and Asia (Çiftçi and Şevkli, 2015; http://www. that the interval type-2 fuzzy TOPSIS method can be utilized
dailymail.co.uk). effectively in the decision-making process of new route analysis.
Hub and spoke network systems provide important benefits for Furthermore, a sensitivity analysis has been conducted by changing
network carriers (Brueckner and Spiller, 1994; Caves et al., 1984; the priorities of the criteria for this model.
Jaillet et al., 1996; Çiftçi and Şevkli, 2015). The following are some The rest of the paper is organized as the following. Section 2,
of these benefits: aggregation of passenger numbers and providing provides an overview of the relevant work. Section 3, presents the
economies of density; minimizing the number of routes needed to basic steps of the interval-valued type-2 fuzzy and the proposed
connect cities in a network; increasing the demand for frequent integrated method based on interval type-2 fuzzy TOPSIS. Section 4,
flights; combining the activities of staff, maintenance, and opera- presents a case study of new route selection between Turkey-North
tions; decreasing costs; and raising customer loyalty with airport American region destinations. Section 5 discusses the results.
domination. Finally, Section 6 concludes this study.
The airline company subject to consideration in this paper also
utilizes hub and spoke network systems. This company feeds its 2. Literature review
network with long routes. Besides, the company inaugurates the
operation of either two or three new long routes each year (due to Multi Criteria Decision Making, is the process of making the best
fleet expansion). While opening new routes, the effect of the new decision from a group of decisions involving more than one crite-
destinations in terms of passenger flow on current route network is rion and decision maker. In fact, MCDM is the process of decision
also considered. For example, it is important to examine whether maker's choosing between more than one alternative according to
opening a new route to North American region also increases the certain criteria (Mardani et al., 2015a, 2015b; 2015c; Deveci et al.,
passenger demand of other regions like South American, African or 2015; Yavuz and Deveci, 2014; Demirel et al., 2016).
Asia Pacific regions. A vast number of very successful fuzzy MCDM have been applied
Fuel expenses stand out as the most important and volatile cost to a variety of problems, like the selection of outsourcing providers
item for airline companies. Apart from this, passenger costs (meal, in the airline industry (Liou and Chuang, 2010), strategic alliance
inflight entertainment etc.), station costs, gate costs and parking partner selection (Liou et al., 2011; Liou, 2012; Prakash and Barua,
costs are among the variable costs as the company realizes addi- 2015; Garg, 2016), third-party reverse logistics partner selection
tional operations. Pilot and crew salaries, aircraft leasing costs can (Prakash and Barua, 2016a, 2016b), ranking of the airports (Prakash
be regarded as the fixed costs. Finally, administrative costs that are and Barua, 2016c), evaluating service quality of airlines (Chang and
not related with operations includ office rents, marketing and sales Yeh, 2002; Chen et al., 2011) and investigating an efficiency
costs, distribution channel costs, managers and office personnel assessment of airlines (Barros and Wanke, 2015; Wanke et al.,
salaries comprise the indirect costs. By dividing the available seat 2015).
kilometer with the total costs, cost per available seat kilometer can In recent years, interval type-2 fuzzy sets together with multi
be found. Airline companies utilize new and productive fleet criteria decision-making techniques have been applied on various
especially wide-body aircraft for inaugurating long-haul new des- kind of problems. Mendel et al. (2006) present a novel method that
tinations in order to minimize the effect of fixed and indirect costs makes an interval type-2 fuzzy logic system much more accessible
and increase profitability. to all readers of this study. Wang et al. (2012) investigate multi-
Network development can be helpful in order to provide cred- attribute group decision making models under interval type-2
ible analysis of profitable new route opportunities for airline de- fuzzy sets. Gong (2013) presents a fuzzy multi-attribute group de-
cision makers. For this reason, the airlines are looking for viable and cision making method based on ınterval type-2 fuzzy sets and
profitable markets (http://www.aviationeconomics.com). applications to global supplier selection. And global supplier se-
In recent years, new route analysis has become an important lection problem is given to illustrate the feasibility and effective-
measure of success and it informs strategic decisions. However, this ness of the proposed method. Hu et al. (2013) proposes a new
decision is usually complicated and is based on many consider- approach based on possibility degree to solve MCDM problems in
ations. Many qualitative and quantitative criteria need to be taken which the criteria value takes the form of interval type-2 fuzzy
into account at the same time. Fuzzy multi-criteria decision-mak- number. This method is applied to a case study on the overseas
ing methods have been developed as a solution to deal with these minerals investment for one of the largest multi-species nonferrous
problems. These methods reflect a new aspect of the selection metals companies in China and the results show the feasibility of
made by decision-makers and they evaluate numerous criteria the method. An overview of previous work on relevant MCDM
simultaneously. The airline company examined in this study plans studies is provided in Table 1.
to incorporate three new wide-body aircraft into its fleet. With
these new three aircraft, due to network structure, one far east and 3. Methodology
one westbound trip can be scheduled rotationally. For this reason,
five alternative destinations from North American region are The proposed methodology consists of three basic stages: (1)
deemed available. Identification of the criteria, alternatives and linguistic variables to
The objectives of this study are to: (1) Construct a methodology be used in the model (2) Analysis of method using these selected
for destination selection problem for network carrier airlines (2) To criteria, alternatives and linguistic variables (3) Ranking the alter-
identify a real life airline industry case (3) To provide a solution to natives using interval type-2 fuzzy TOPSIS. The schematic diagram
this case by TOPSIS method, one of the multi-criteria decision- of the proposed methodology for the selection of new route is
making methods, which has been integrated with the interval type- shown in Fig. 1. The stages are as follows:
M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99 85

Table 1
An overview of some previous studies on multi-criteria decision making selection problems.

Author (year) Subject area Interval-valued TOPSIS VIKOR Servqual GRA Choquet AHP
fuzzy integral

Lee and Chen (2008) Evaluation of car alternatives Type 2 x


Ashtiani et al. (2009) Personnel selection problem for telecommunication company Type 2 x
Chen and Lee (2010a) Personnel selection problem for software company Type 2 x
Kuo (2011) Evaluation of airline service quality Type 2 x x
Chen (2012) Computational experiment Type 2 x
Celik et al. (2013) Evaluation of public transportation service quality Type 2 x x
Tan and Chen (2013) Evaluation of investment company Type 2 x x
Celik et al. (2014) Evaluation of customer satisfaction Type 2 x x
Cevik Onar et al. (2014) Strategic decision selection Type 2 x x
Mokhtarian et al. (2014) Facility location selection problem Type 2 x
Dymova et al. (2015) Personnel selection problem for software company Type 2 x
Kahraman et al. (2014) Supplier selection problem Type 2 x
Kilic and Kaya (2015) Evaluation of investment project Type 2 x x
Liao (2015) Material selection Type 2 x
Qin et al. (2015) Project investment evaluation problem Type 2 x
Ghorabaee (2016) Robot selection Type 2 x

Fig. 1. The steps of the proposed methodology.


86 M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99

Stage 1: Form the fuzzy model using selected criteria, location


alternatives and a team of decision makers. Also fuzzy weights of Z Z
e
~¼ 1
each criterion and alternative are computed. A ;
ðx; uÞ
Stage 2: Determine the fuzzy decision matrix, normalized and xεX uεJx
weighted etc.
Stage 3: Rank each alternative based on the outcome from Stage where Jx 4½0; 1.
2.
Definition 3. (Mendel et al., 2006). Type-1 membership functions
The steps of this methodology are constructed with inspiration
are comprised of the upper membership function and the lower
by Lee and Chen (2008) in the literature. The primary logic of in-
membership function of an interval type-2 fuzzy set, respectively.
terval type-2 Fuzzy TOPSIS is the determination of positive ideal
In this paper, we present a method to use interval type-2 fuzzy sets
and negative ideal distances of the alternatives and according to
for dealing with fuzzy multiple attributes group decision-making
that making a ranking between the alternatives. As a result, finding
problems. In these problems, the reference points and the heights
the best alternative is aimed with this MDCM technique.
of the upper and the lower membership functions of interval type-2
fuzzy sets are utilized to characterize interval type-2 fuzzy sets.
Fig. 2 shows a trapezoidal interval type-2 fuzzy set

3.1. Interval Type-2 fuzzy sets  U L   U  U 


e
~ ¼ A~ ; A
~ ¼ aU ; aU ; aU ; aU ; H A ~ ;H A ~
A i i i i1 i2 i3 i4 1 i 2 i ;
In this section, we present some basic definitions of type-2 fuzzy   L  L 
aLi1 ; aLi2 ; aLi3 ; aLi4 ; H1 A~ ;H A ~
sets and interval type-2 fuzzy sets from Lee and Chen (2008), Chen i 2 i
and Lee (2010a, 2010b), Mendel et al. (2006), Ashtiani et al. (2009),
Kilic and Kaya (2015), Celik et al. (2013), Gong (2013), Hu et al. (Lee & Chen, 2008), where A ~ U and A
~ L are type-1 fuzzy sets,
i i
(2013), Mokhtarian et al. (2014), Wang et al. (2012). aU U U U L L L L
i1 ; ai2 ; ai3 ; ai4 ; ai1 ; ai2 ; ai3 and ai4 are the reference points of the
e
~ in the universe of discourse X e ~U
~ , H ðA
Definition 1. A type-2 fuzzy set A interval type-2 fuzzy set A i j i Þ denotes the membership value
can be represented by a type-2 membership function me , shown as of the element aU
A~ iðjþ1Þ in the upper trapezoidal membership func-
follows (Mendel et al., 2006):
~ U , 1  j  2, H ðA
~L
tion A i j i Þ denotes the membership value of the
 
e  ~L,
element aLiðjþ1Þ in the lower trapezoidal membership function A

A ðx; uÞ; me ðx; uÞ  cx ε X; cu ε Jx 4½0; 1; 0 i
~
A
 ~LÞ
1  j  2, Hj ðAi
 me ðx; uÞ  1
~
A  U  U
~ L ; 1  j  2; H A
A ~ ε½0; 1; H A ~ ε ½0; 1;
i 1 i 2 i
where Jx denotes an interval in [0, 1]. Moreover, the type-2 fuzzy set  L  L
e
~ also can be represented as follows (Mendel et al., 2006): H1 A ~ ε ½0; 1; H A ~ ε ½0; 1; and 1  i  n
A i 2 i

Z Z
e
~¼ ðx; uÞ
A me~ ;
A ðx; uÞ Some definitions will be given below (Lee and Chen, 2008):
xεX uεJx
Definition 4. The addition operation can be carried out between
where Jx 4½0; 1 and ∬ denotes union over all admissible x and u. the trapezoidal interval type-2 fuzzy sets
e
~ be a type-2 fuzzy set in the universe of
Definition 2. Let A
discourse X denoted by the type-2 membership function me . If all  U L    U  U 
e
~ ¼ A~ ; A
~ ¼ aU ; aU ; aU ; aU ; H A ~ ;H A ~
~
A A 1 1 1 11 12 13 14 1 1 2 1 ;
e
~ is counted as an interval type-2 fuzzy set. An   L  L 
me~ ðx; uÞ ¼ 1, then A ~ ;H A ~
A aL11 ; aL12 ; aL13 ; aL14 ; H1 A 1 2 1
e
~ can be regarded as a special case of a
interval type-2 fuzzy set A
type-2 fuzzy set, indicated as the following (Mendel et al., 2006): and

~ U and the lower trapezoidal membership function A e


~ L of the interval type-2 fuzzy set A
~ (Lee and Chen, 2008).
Fig. 2. The upper trapezoidal membership function A i i i
M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99 87

interval type-2 fuzzy set indicated in Fig. 2, where


 U L    U  U 
e
~ ¼ A~ ; A
~ ¼ aU ; aU ; aU ; aU ; H A ~ ;H A ~
A 2 2 2 21 22 23 24 1 2 2 2 ;
  L  L   U L
~ ;H A ~ e
~ ¼ A ~ ; A ~
aL21 ; aL22 ; aL23 ; aL24 ; H1 A 2 2 2
A i i i
  U  U
¼ aU U U U ~ ~
 U L   U L   i1 ; ai2 ; ai3 ; ai4 ; H1 Ai ; H2 Ai
e
~ 4A e
~ ¼ A ~ ; A
~ 4 A ~ ; A ~ ¼    L  L 
A 1 2 aU U U U U
11 þ a21 ; a12 þ a22 ; a13 ~ ;H A ~
1 1 2
  U
2
 U    U  ; aLi1 ; aLi2 ; aLi3 ; aLi4 ; H1 Ai 2 i
þaU U U ~ ~ ~ ;
23 ; a14 þ a24 ; min H1 A1 ; H1 A2 ; min H2 A 1
 U  
~
H2 A2 L L L L L L L L
; a11 þ a21 ; a12 þ a22 ; a13 þ a23 ; a14 þ a24 ;
  L  L    L  L  e
~ Þ of the trapezoidal interval type-2
~ ~ ~ ;H A ~ The ranking value Rank ðA i
min H1 A1 ; H1 A 2 ; min H2 A 1 2 2 e
~
fuzzy set Ai is defined as follows (Lee and Chen, 2008):

   U  L  U  L
e
~ ¼M A ~ þM A ~ þM A ~ þM A ~
Definition 5. The subtraction operation can be carried out be- Rank A i 1 i 1 i 2 i 2 i
tween the trapezoidal interval type-2 fuzzy sets  U  L 1   U   L
~ þM A
þM3 A ~  ~ þS A
S A ~
3
   U L   i i
4 1 i 1 i
e
~ .A e
~ ¼ A ~U ; A
~L . A ~ ; A
~ ¼  U  L  U  L  U
A aU U U U U
11  a24 ; a12  a23 ; a13
1 2 1 1 2
  U
2
 U    U ~ þS A
þS A ~ þS A ~ þS A ~ þS A ~
2 i 2 i 3 i 3 i 4 i
aU U U ~ ~ ~ ;
22 ; a14  a21 ; min H1 A1 ; H1 A2 ; min H2 A 1    U  L  U  L
 U   ~L ~ þH A ~ þH A ~ þH A ~
~ L L L L L L L L þS4 A þ H1 A
H2 A2 ; a11  a24 ; a12  a23 ; a13  a22 ; a14  a21 ; i i 1 i 2 i 2 i
  L  L    L  L 
min H A ~ ;H A ~ ; min H A ~ ;H A ~
1 1 1 2 2 1 2 2 ~ j Þ denotes the average of the elements aj and aj
where Mp ðA i ip iðpþ1Þ

 j  aj þ aj
Definition 6. The multiplication operation can be carried out ~ ¼ ip
Mp A
iðpþ1Þ
; 1  p  3;
i
between the trapezoidal interval type-2 fuzzy sets 2

0   U  U    U  U  1
 U L  U L aU xaU ; aU xaU ; aU x aU ; aU xaU ; min H1 A ~ ;H A ~ ; min H2 A ~ ;H A ~ ;
e
~ 5Ae
~ ¼ A~ ; A ~ ; A
~ 5 A ~ ¼@  11 21 12 22 13 23 14 24 1
  L
1 2
 L    L
1 2 2
 L  A
A 1 2 1 1 2 2
~ ;H A
aL11 xaL21 ; aL12 xaL22 ; aL13 x aL23 ; aL14 x aL24 ; min H1 A ~ ; min H2 A~ ;H A ~
1 1 2 1 2 2

vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u 0 12
 j u
u1 X
qþ1
1 X
qþ1
Definition 7. The arithmetic operations can be done between the ~ ¼t @a 
Sq Ai
j
ik
aik A
j
trapezoidal interval type-2 fuzzy sets 2 2
k¼q k¼q
    U  U 
e
~ ¼ ~U ; ~L ~ ~ ~ j Þ denotes the standard deviation of the elements
A 1 A1 A 1 ¼ aU U U U
11 ; a12 ; a13 ; a14 ; H1 A1 ; H2 A1 ; 1  p  3; S4 ðA i
  L  L  j j j j
ai1 ; ai2 ; ai3 ; ai4 ,
~ ;H A
aL11 ; aL12 ; aL13 ; aL14 ; H1 A ~
1 2 1 vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u !2
 j  u1 X 4
1X 4
and the crisp value k is defined as follows: ~
S4 A i ¼ t j
aik  aikj
4 4
k¼1 k¼1
0   U  1 
xaU xaU xaU ~U ; H A ~
xaU ~ j Þ denotes the membership value of the element aj
e k
~ ¼@ 
k 11 ; k k12 ; 13 ;
A 1 2 1 ;
14 ; H1 Hp ðA in
 L  A
i iðpþ1Þ
kA 1  L
L L L L ~ ~ ~ j ; 1  p  2; jЄfU; Lg, and
k xa11 ; k xa12 ; k xa13 ; k xa14 ; H1 A1 ; H2 A1 the trapezoidal membership function A i
1in

0  U  U  1 3.2. The proposed TOPSIS method based on interval Type-2 fuzzy


1 U 1 U 1 U 1 U ~ ;H A ~
e B xa11 ; xa12 ; xa13 ; xa14 ; H1 A ; C numbers
~
A B k k k k 1 2 1
C
1
¼B   L  L  C
k @ 1 L 1 L 1 L 1 L A
~ ;H A
xa11 ; xa12 ; xa13 ; xa14 ; H1 A ~ The interval type-2 fuzzy TOPSIS methods, used in the solution
1 2 1
k k k k of the new route selection problem is a multi-criteria decision
making method created by adapting the type-2 fuzzy sets to the
where k > 0. TOPSIS method. The TOPSIS (Technique for Order Preference by
Similarity To An Ideal Solution) method, which was first put for-
Definition 8. Lee and Chen (2008) presented the method of
ward in 1981 as a solution to multi-criteria decision making
e
~ be an
ranking values of trapezoidal interval type-2 fuzzy sets. Let A i methods (Hwang and Yoon, 1981), is based on the idea of
88 M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99

determining the alternatives with the shortest geometric distance and


from the positive ideal solution (PIS) and the longest geometric
distance from the negative ideal solution (NIS), and establishing a
x1 x2 … xm h
ranking according to this (Chen, 2000). In the TOPSIS method, an y i
(4)
Wz ¼ w e
~i ¼ w e y
~ 1; e y
~ 2; e
~m
y
index called distance to PIS and NIS is defined. With this definition,
1xm
w …; w
the method chooses the alternative with the maximum similarity
to the ideal solution (Yoon and Hwang, 1995). 0 1
e~x 4 e~x …4e~x
1 2 k

iA e
Step 1. Determine criteria and alternatives. Let us assume that where e
xij ¼ @ i
~ i
k
~i is an interval type-2 fuzzyset or crisp
;a
there is a set A of alternatives, where A ¼ {a1, a2, …, an}, and assume
that there is a set X of criteria, X ¼ {x1, x2, …, xm}. number, 1  i  m, 1  j  n, 1  y  k and denotes the number of
Step 2. Form a committee of decision-makers. Assume that decision-makers.where e ~
x c i,j and w~ ; i ¼ 1,2, ….,m are linguistic
ij i
there are k decision-makers DM1, DM2, …, and DMk.
variables which can be described by interval type-2 fuzzy numbers,
Step 3. Choose linguistic variables for evaluating criteria and
e
~
x ¼ ðaU ; aU ; aU ; aU Þ and e
~ ¼ ðwU ; wU ; wU ; wU Þ,
w
alternatives. The linguistic variables used for determining the ij i1 i2 i3 i4 i i1 i2 i3 i4
criteria weights, significance degrees of the alternatives and the e
~ e
~ i ¼ ðwLi1 ; wLi2 ; wLi3 ; wLi4 Þ.
xij ¼ ðaLi1 ; aLi2 ; aLi3 ; aLi4 Þ and w
corresponding fuzzy numbers are indicated in Tables 2 and 3. Step 6. Normalize the fuzzy decision matrix.
Step 4. Fuzzy weights of each criterion and alternative are
calculated using the equations numbered (1) and (2). “k” in the h i
e
~¼ e
~r ij
equation corresponds to the number of decision makers. R i ¼ 1; 2; …; m ; j ¼ 1; 2; …; n (5)
mxn
2 3
~ 1 Xk
~ y where B and C are the sets of benefit criteria and cost criteria,
~i ¼ 4
w ~ 5;
w i ¼ 1; 2; …:; m (1)
k y¼1 i respectively as in the following:

2 3 !
aU U U U  U  U
1 X
k e
~r ij ¼ i1 ai2 ai3 ai4 ~ ;H A ~
~
xij ¼ 4
~ ~y
x 5;
~ j ¼ 1; 2; …:; n (2) ; ; ; ; H 1 A i 2 i ;
k y¼1 ij aUi*
aU
i*
aU
i*
aU
i*
!! (6)
aLi1 aLi2 aLi3 aLi4  L  L
~
; U ; U ; U ; H1 A1i ; H2 Ai ~
e
~
xij is the degree of “i”th criterion according to “j”th alternative and aU ai* ai* ai*
i*
e
~ i is the significance weight of “i”th criterion.
w
Step 5. Structure the fuzzy decision matrix. The fuzzy decision
!
matrix is created using the equations numbered (3) and (4). The
aU U U U  U  U
e e
~ and the criteria (w)
~ is e
~r ij ¼ i ai ai ai ~ ~
fuzzy decision matrix for the alternatives (A) ; U ; U ; U ; H 1 Ai ; H 2 Ai ;
constructed as follows: aU a a a
i4 i3 i2 i1
!! (7)
2 3 aU U U U  L  L
i ai ai ai ~ ;H A ~
ez
ez
ez ; ; ; ; H A
x1 6 ~
x11 ~
x12 / ~
x1n aLi4 aLi3 aLi2 aLi1
1 1i 2 i
  7
x2 6 7
z z z
6 e
~
x11 e
~
x22 / e
~
x2n 7
xzij
Dz ¼ ~ ¼ (3)
mxn « 6
4 « « « « 7
5
xm e z z z
~ e
~ … e
~ aU U
aU U
i* ¼ maxaij ; j ε Bðbenefit criteriaÞ and i ¼ minaij ;
xm1 xm2 xmn
i i
j ε Cðcost criteriaÞ:

Table 2
e
~ : Normalized fuzzy decision matrix
Linguistic variables for the importance weight of each criterion (Chen and Lee, R
2010a,b). aUij : Maximum value of the fourth component in one column in
Linguistic variables Interval type-2 fuzzy sets fuzzy decision matrix
e
~r ij : Normalized values obtained by dividing each value in fuzzy
Very low (VL) ((0, 0, 0, 0.1; 1, 1), (0, 0, 0, 0.05; 0.9, 0.9))
Low (L) ((0, 0.1, 0.1, 0.3; 1, 1), (0.05, 0.1, 0.1, 0.2; 0.9, 0.9)) decision matrix into cj value.
Medium low (ML) ((0.1, 0.3, 0.3, 0.5; 1, 1), (0.2, 0.3, 0.3, 0.4; 0.9, 0.9))
Medium (M) ((0.3, 0.5, 0.5, 0.7; 1, 1), (0.4, 0.5, 0.5, 0.6; 0.9, 0.9)) Step 7. Structure the weighted normalized matrix.
Medium high (MH) ((0.5, 0.7, 0.7, 0.9; 1, 1), (0.6, 0.7, 0.7, 0.8; 0.9, 0.9))
High (H) ((0.7, 0.9, 0.9, 1; 1, 1), (0.8, 0.9, 0.9, 0.95; 0.9, 0.9))
Very high (VH) ((0.9, 1, 1, 1; 1, 1), (0.95, 1, 1, 1; 0.9, 0.9)) h i
~
~v ¼ ~
~v ; i ¼ 1; 2; …; m ; j ¼ 1; 2; …; n where ~
~vij
ij mxn

¼~ ~
~rij ð$Þw
~i (8)
Table 3
Linguistic variables for the ratings (Chen and Lee, 2010a,b).
where e~vij denotes the weighted normalized interval type-2 fuzzy
Linguistic variables Interval type-2 fuzzy sets
numbers.
Very poor (VP) ((0, 0, 0, 1; 1, 1), (0, 0, 0, 0.5; 0.9, 0.9)) Step 8. Based on Eq. (9) and Eq. (10), calculate the ranking value
Rankðe
~v Þ of the interval type-2 fuzzy set e
Poor (P) ((0, 1, 1, 3; 1, 1), (0.5, 1, 1, 2; 0.9, 0.9))
ij ~v , where 1  j  n.
ij
Medium poor (MP) ((1, 3, 3, 5; 1, 1), (2, 3, 3, 4; 0.9, 0.9))
Fair (F) ((3, 5, 5, 7; 1, 1), (4, 5, 5, 6; 0.9, 0.9)) Step 9. Compute the distance of each alternative from fuzzy
Medium good (MG) ((5, 7, 7, 9; 1, 1), (6, 7, 7, 8; 0.9, 0.9)) positive-ideal solution (FPIS, A*) A ¼ ðv*1 ; v*1 ; …; ; v*m Þ and fuzzy
Good (G) ((7, 9, 9, 10; 1, 1), (8, 9, 9, 9.5; 0.9, 0.9)) negative-ideal solution (FNIS, A) A ¼ ðv  
1 ; v1 ; …; ; vm Þ, respec-
Very good (VG) ((9, 10, 10, 10; 1, 1), (9.5, 10, 10, 10; 0.9, 0.9))
tively as in the following:
M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99 89



8 n  o
d x
< max Rank e CC xj ¼
j
(13)
~vij ; i  j  n ; if fi Є B d xj þ d* xj
* n  o
vi ¼ (9)
: min Rank e
~vij ; i  j  n ; if fi Є C Step 11. The last, rank the alternatives according to their close-
ness coefficients that are between 0 and 1, and finally choose the
and
alternative whose closeness coefficient is adjacent to 1.

8 n  o
< min Rank e~vij ; i  j  n ; if fi Є B 4. Case study
v
i ¼:
n  o (10)
e
min Rank ~vij ; i  j  n ; if fi Є C Medium term capacity planning in airline companies is gener-
ally done by network planning departments. The analysis of new
j Þ of each weighted alternative j ¼ 1, 2, …, n
The distance ðd*j ; d destinations is also carried out by network analysis or network
from the fuzzy positive-ideal solution (FPIS, A ~ * ) and the fuzzy development departments. For these studies, the past commercial
~  ) is computed as follows:
negative ideal solution (FNIS, A performance data (load factor, revenues, cost analysis etc.) of the
current routes are compared with the previous term. The market
trends are also analysed. Feasibility analysis and business cases for
  rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
   2

* the short haul routes are prepared 6e8 months before the sched-
~v0j ; e
d*j xj ¼ d e ~vij ¼ Rank e ~vij  v*i (11)
uling seasons and 10e12 months for the long-haul routes. Case
studies are not only done for capacity demands of current routes
but also for new route feasibilities. In new route studies, on the
  rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
   2 other hand, potential revenues, profitability and potential market


d ¼ e
~v e
~ Rank e ~vij  v
j x j d 0j ij ¼
; v i
(12) shares are analysed continuously, and alternative scenarios are
prepared for sensitivity analysis. These scenarios include new
Step 10. Calculate the closeness coefficient CCðxj Þ of each destination selection among alternative destinations, decision of
alternative. weekly frequency and fleet assignment case studies.

Fig. 3. The decision hierarchy of the new route analysis problem.


90 M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99

Fig. 4. Map of North American region showing the alternative airports from Istanbul.

With the interval type-2 fuzzy TOPSIS method, new route at least 2 million international passengers per year (IADS, 2014).
analysis with MCDM is the subject matter of this study. It is Five expert decision makers (DM) have been determined and
intended to evaluate potential destinations and to choose the most questionaries are conducted with these decision makers, five
appropriate alternative. As mentioned above, the airline company alternative destinations and eleven criteria to evaluate these al-
examined in this study plans to incorporate three new wide-body ternatives have been determined. In Fig. 3 the general outline of
aircraft into its fleet. With these new three aircraft, due to this problem is shown. The alternative airports are shown in Fig. 4.
network structure, one far east and one westbound trip can be In this study, eleven different criteria are employed, identified
scheduled rotationally. In this study, five alternative destinations and synthesized for the new route analysis problem. Each criterion
from North American region are deemed available by network is presented and explained in Table 4. These criteria are demand
planning experts. These five alternative airports are selected for related and employed in the industry to determine and analyze the
case study because they are among top 25 US busiest airports in air travel demand for a certain route (Air Travel Demand, 2008).
terms of international passenger figures of year 2014 and they have Airlines analyze historic demand data using these criteria from

Table 4
Evaluation criteria and descriptions.

Criteria Definition Criteria


type

C1: City Population “City Population” is an important parameter for opening a flight line in a city. Benefit
C2: GDP and Trade Relations GDP (Gross Domestic Product); is a monetary measure of the value of all final goods and services produced in a period in a Benefit
country (http://climatepolicyinfohub.eu).The number of traveling passengers by plane is proportional to the size of the
economy.
C3: Potential Passengers to Indicates the number of passengers flying to and from (omni directional) any destination. Benefit
Turkey
C4: Number of Transit Indicates the number of passengers going to and coming from another city via a connecting flight. For example, passengers Benefit
Passengers who are planning to go to Hong Kong from London, flying to Dubai with the London-Dubai flight, then taking second flight
Dubai-Hong Kong. Passengers traveling in this way are called transit passengers.
Note: In our study, the number of passengers traveling from any point in North American region to any other eastern points in terms
of geographical location over the last year is regarded as transit passenger number. Big markets (transit passenger number) have
positive impact on the launch of new route.
C5: Number of Business Passenger profiles are very important in those passengers with business purposes being numerous and these passengers fly Benefit
Passengers very often, is an advantage to the airline company in terms of revenue.
C6: Airport Unit Passenger It shows the average unit revenue of passengers fly to/from the airport. (For instance, East Midland Airport is mostly used by Benefit
Revenue charters and low cost passengers whereas Heathrow is preferred by business people. The more airport unit revenue leads to
more high yield).
C7: Competitiveness Index If there are many competitors in a market, the figure will be higher. If there is only one company, it will be 1. Benefit
C8: Number of Competitors It is frequency of flights, which means the number of flights with in a specified time. Cost
Flights
C9: Distance and Cost As a natural outcome of the distance covered by the planes increasing; the costs regarding fuel, cabin attendants also increase Cost
thus leading to an overall increase in travel cost (Ertürk, 2011).
C10: Touristic and Business A high touristic and business customer number is another advantage to the airline company in terms of revenue. Benefit
Customer Flow
C11: Seasonality Seasonality being high in an X destination may turn out to be a disadvantage in some cases. A sustainable capacity usage is Cost
another important aspect.
M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99 91

their networks perspective, estimate their market shares, make Number of business passengers and number of potential
feasibility studies. There are other criteria that are considered while transfers or transit passengers are added values of the long-haul
opening a new route such as exchange rates, countries's social and routes. When demand decreases occur on local passengers, con-
economic conditions, security and safety, risk factors, insurance necting the passengers can compensate this decrease. However
policy etc. However these are non-demand related factors that transit passengers have usually low yields because a connecting
affect the selection of a new destination. itinerary fare is usually lower than the sum of the local fares from
The first and most basic criteria for a new route selection is the each segments of the connecting passenger itinerary (Bratu, 1998).
population of the city. Population shows the potential demand and The number of business passengers is the source of high yield and it
accessibility of airports from a specific location. According to the has great impact on route profitability.
IATA Air Travel Demand briefing, population has a direct effect on Unit passenger revenues have similar affects on route profit-
the size of a market (Air Travel Demand, 2008). However it is not ability. They are identified as a benefit because higher expected unit
the only thing that affects market sizes. There is a very high cor- revenues generate higher revenue under the same demand. While
relation between higher GDP per capita and penetration of air competition on a route increases the market size figures and
travel (http://centreforaviation.com). Potential passenger to/from stimulates traffic and passenger figures, it causes decline in unit
Turkey, local passengers, and trade relations between two countries passenger revenues. Also competitors’ products, the number of
also have impact on demand for routes. This case presents an alternative frequencies, departure times, aircraft types, catering
example for local passengers as the most important revenue and service quality have great impact on passenger choice and
generator for a monopoly route. The airline company will have all of market share.
the revenues and higher demand leads to more revenue from new Sustainable and high demand from business and touristic pas-
route. sengers are important to gain market share. High demand fluctu-
ations between summer and winter cause low utilization of
aircrafts and increase the cost of aircraft ownership per airline seat
kilometers.
Table 7
Fuzzy weights for criteria. 4.1. The proposed method computations
Criteria Weights
The linguistic assessments for the eleven criteria are determined
C1 ((0.58; 0.76; 0.76; 0.88; 1;1),(0.67; 0.76; 0.76; 0.82; 0.9; 0.9))
C2 ((0.54; 0.72; 0.72; 0.86; 1;1),(0.63; 0.72; 0.72; 0.79; 0.9; 0.9)) by the committee using rating scales (see Table 2), which also
C3 ((0.74; 0.9; 0.9; 0.98; 1;1),(0.82; 0.9; 0.9; 0.94; 0.9; 0.9)) evaluate the five alternatives (airports) for each of the 11 criteria
C4 ((0.62; 0.82; 0.82; 0.96; 1;1),(0.72; 0.82; 0.82; 0.89; 0.9; 0.9)) (using rating scales of Table 3). The linguistic assessments for the
C5 ((0.36; 0.54; 0.54; 0.72; 1;1),(0.45; 0.54; 0.54; 0.63; 0.9; 0.9)) criteria and alternatives are presented in Tables 5 and 6 of the
C6 ((0.58; 0.76; 0.76; 0.9; 1;1),(0.67; 0.76; 0.76; 0.83; 0.9; 0.9))
C7 ((0.74; 0.9; 0.9; 0.98; 1;1),(0.82; 0.9; 0.9; 0.94; 0.9; 0.9))
appendix.
C8 ((0.66; 0.84; 0.84; 0.96; 1;1),(0.75; 0.84; 0.84; 0.9; 0.9; 0.9)) The fuzzy weights of the 11 criteria are presented in Table 7. The
C9 ((0.78; 0.94; 0.94; 1;1; 1),(0.86; 0.94; 0.94; 0.97; 0.9; 0.9)) fuzzy weights (w ~ j ) for each criterion are computed by using Eq. (1).
C10 ((0.58; 0.76; 0.76; 0.9; 1;1),(0.67; 0.76; 0.76; 0.83; 0.9; 0.9)) For criterion C1 ‘‘City Population’’, the fuzzy weight is given by w ~i ¼
C11 ((0.5; 0.7; 0.7; 0.86; 1;1),(0.6; 0.7; 0.7; 0.78; 0.9; 0.9)) ~ i1 ; w
~ i2 ; w
~ i3 where
w

Table 8
The fuzzy decision matrix.

Criteria A1 Airport A2 Airport A3 Airport A4 Airport A5 Airport

C1 ((6.6; 8.4; 8.4; 9.6; 1;1), (7.5; ((3.4; 5.4; 5.4; 7.2; 1;1), (4.4; ((7.8; 9.2; 9.2; 9.8; 1;1), (8.5; ((3.4; 5.2; 5.2; 6.8; 1;1), (4.3; 5.2; ((6.2; 8.2; 8.2; 9.6; 1;1), (7.2;
8.4; 8.4; 9;0.9; 0.9)) 5.4; 5.4; 6.3; 0.9; 0.9)) 9.2; 9.2; 9.5; 0.9; 0.9)) 5.2; 6;0.9; 0.9)) 8.2; 8.2; 8.9; 0.9; 0.9))
C2 ((5.4; 7.4; 7.4; 9;1; 1), (6.4; ((3.4; 5.4; 5.4; 7.4; 1;1), (4.4; ((5.8; 7.6; 7.6; 8.8; 1;1), (6.7; ((2; 3;3; 4.4; 1;1), (2.5; 3;3; 3.7; 0.9; ((2.2; 4.2; 4.2; 6.2; 1;1), (3.2;
7.4; 7.4; 8.2; 0.9; 0.9)) 5.4; 5.4; 6.4; 0.9; 0.9)) 7.6; 7.6; 8.2; 0.9; 0.9)) 0.9)) 4.2; 4.2; 5.2; 0.9; 0.9))
C3 ((7.4; 9;9; 9.8; 1;1), (8.2; 9;9; ((2.8; 4.4; 4.4; 6;1; 1), (3.6; ((6.6; 8.4; 8.4; 9.6; 1;1), (7.5; ((2.5; 3.25; 3.25; 4.25; 1;1), (2.88; ((3.4; 5.4; 5.4; 7.4; 1;1), (4.4;
9.4; 0.9; 0.9)) 4.4; 4.4; 5.2; 0.9; 0.9)) 8.4; 8.4; 9;0.9; 0.9)) 3.25; 3.25; 3.75; 0.9; 0.9)) 5.4; 5.4; 6.4; 0.9; 0.9))
C4 ((4.2; 6.2; 6.2; 8.2; 1;1), (5.2; ((5.8; 7.6; 7.6; 8.8; 1;1), (6.7; ((6.6; 8.4; 8.4; 9.4; 1;1), (7.5; ((7.8; 9.4; 9.4; 10; 1;1), (8.6; 9.4; ((8; 9.5; 9.5; 10; 1;1), (8.75;
6.2; 6.2; 7.2; 0.9; 0.9)) 7.6; 7.6; 8.2; 0.9; 0.9)) 8.4; 8.4; 8.9; 0.9; 0.9)) 9.4; 9.7; 0.9; 0.9)) 9.5; 9.5; 9.75; 0.9; 0.9))
C5 ((7; 8.8; 8.8; 9.8; 1;1), (7.9; ((5; 6.8; 6.8; 8.2; 1;1), (5.9; ((6.2; 8;8; 9;1; 1), (7.1; 8;8; ((3.4; 5.4; 5.4; 7.4; 1;1), (4.4; 5.4; ((5.8; 7.6; 7.6; 8.8; 1;1), (6.7;
8.8; 8.8; 9.3; 0.9; 0.9)) 6.8; 6.8; 7.5; 0.9; 0.9)) 8.5; 0.9; 0.9)) 5.4; 6.4; 0.9; 0.9)) 7.6; 7.6; 8.2; 0.9; 0.9))
C6 ((3.8; 5.8; 5.8; 7.6; 1;1), (4.8; ((4.2; 6;6; 7.4; 1;1), (5.1; 6;6; ((7.8; 9.2; 9.2; 9.8; 1;1), (8.5; ((2.8; 4.4; 4.4; 6;1; 1), (3.6; 4.4; 4.4; ((5.4; 7.4; 7.4; 9;1; 1), (6.4;
5.8; 5.8; 6.7; 0.9; 0.9)) 6.7; 0.9; 0.9)) 9.2; 9.2; 9.5; 0.9; 0.9)) 5.2; 0.9; 0.9)) 7.4; 7.4; 8.2; 0.9; 0.9))
C7 ((5; 7;7; 8.6; 1;1), (6; 7;7; 7.8; ((4; 5.8; 5.8; 7.4; 1;1), (4.9; ((6.4; 7.8; 7.8; 8.6; 1;1), (7.1; ((5.8; 7.8; 7.8; 9.2; 1;1), (6.8; 7.8; ((6.6; 8.4; 8.4; 9.6; 1;1), (7.5;
0.9; 0.9)) 5.8; 5.8; 6.6; 0.9; 0.9)) 7.8; 7.8; 8.2; 0.9; 0.9)) 7.8; 8.5; 0.9; 0.9)) 8.4; 8.4; 9;0.9; 0.9))
C8 ((3; 5;5; 6.8; 1;1), (4; 5;5; 5.9; ((5.4; 7.2; 7.2; 8.6; 1;1), (6.3; ((5.4; 7.4; 7.4; 9;1; 1), (6.4; ((1.8; 3.2; 3.2; 4.8; 1;1), (2.5; 3.2; ((3.4; 5.4; 5.4; 7.4; 1;1), (4.4;
0.9; 0.9)) 7.2; 7.2; 7.9; 0.9; 0.9)) 7.4; 7.4; 8.2; 0.9; 0.9)) 3.2; 4;0.9; 0.9)) 5.4; 5.4; 6.4; 0.9; 0.9))
C9 ((6.6; 8.4; 8.4; 9.4; 1;1), (7.5; ((3.8; 5.8; 5.8; 7.4; 1;1), (4.8; ((5.8; 7.8; 7.8; 9.2; 1;1), (6.8; ((6.6; 8.2; 8.2; 9.2; 1;1), (7.4; 8.2; ((6.2; 8.2; 8.2; 9.6; 1;1), (7.2;
8.4; 8.4; 8.9; 0.9; 0.9)) 5.8; 5.8; 6.6; 0.9; 0.9)) 7.8; 7.8; 8.5; 0.9; 0.9)) 8.2; 8.7; 0.9; 0.9)) 8.2; 8.2; 8.9; 0.9; 0.9))
C10 ((3;5;5; 7;1; 1), (4;5;5; 6;0.9; ((2.2; 3.8; 3.8; 5.6; 1;1), (3; ((8.2; 9.6; 9.6; 10; 1;1), (8.9; ((3.8; 5.8; 5.8; 7.6; 1;1), (4.8; 5.8; ((5;7;7; 8.6; 1;1), (6;7;7; 7.8;
0.9)) 3.8; 3.8; 4.7; 0.9; 0.9)) 9.6; 9.6; 9.8; 0.9; 0.9)) 5.8; 6.7; 0.9; 0.9)) 0.9; 0.9))
C11 ((5; 7;7; 8.4; 1;1), (6; 7;7; 7.7; ((1.8; 3.4; 3.4; 5.2; 1;1), (2.6; ((6.2; 8.2; 8.2; 9.6; 1;1), (7.2; ((6.6; 8.4; 8.4; 9.6; 1;1), (7.5; 8.4; ((5.8; 7.8; 7.8; 9.2; 1;1), (6.8;
0.9; 0.9)) 3.4; 3.4; 4.3; 0.9; 0.9)) 8.2; 8.2; 8.9; 0.9; 0.9)) 8.4; 9;0.9; 0.9)) 7.8; 7.8; 8.5; 0.9; 0.9))
92 M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99

Table 9
The fuzzy normalized decision matrix.

Criteria A1 Airport A2 Airport A3 Airport A4 Airport A5 Airport

C1 ((0.67; 0.86; 0.86; 0.98; 1;1), ((0.35; 0.55; 0.55; 0.73; 1;1), ((0.8; 0.94; 0.94; 1;1; 1), (0.87; ((0.35; 0.53; 0.53; 0.69; 1;1), ((0.63; 0.84; 0.84; 0.98; 1;1),
(0.77; 0.86; 0.86; 0.92; 0.9; 0.9)) (0.45; 0.55; 0.55; 0.64; 0.9; 0.94; 0.94; 0.97; 0.9; 0.9)) (0.44; 0.53; 0.53; 0.61; 0.9; (0.73; 0.84; 0.84; 0.91; 0.9;
0.9)) 0.9)) 0.9))
C2 ((0.6; 0.82; 0.82; 1;1; 1), (0.71; ((0.38; 0.6; 0.6; 0.82; 1;1), ((0.64; 0.84; 0.84; 0.98; 1;1), ((0.22; 0.33; 0.33; 0.49; 1;1), ((0.24; 0.47; 0.47; 0.69; 1;1),
0.82; 0.82; 0.91; 0.9; 0.9)) (0.49; 0.6; 0.6; 0.71; 0.9; 0.9)) (0.74; 0.84; 0.84; 0.91; 0.9; (0.28; 0.33; 0.33; 0.41; 0.9; (0.36; 0.47; 0.47; 0.58; 0.9;
0.9)) 0.9)) 0.9))
C3 ((0.76; 0.92; 0.92; 1;1; 1), (0.84; ((0.29; 0.45; 0.45; 0.61; 1;1), ((0.67; 0.86; 0.86; 0.98; 1;1), ((0.26; 0.33; 0.33; 0.43; 1;1), ((0.35; 0.55; 0.55; 0.76; 1;1),
0.92; 0.92; 0.96; 0.9; 0.9)) (0.37; 0.45; 0.45; 0.53; 0.9; (0.77; 0.86; 0.86; 0.92; 0.9; (0.29; 0.33; 0.33; 0.38; 0.9; (0.45; 0.55; 0.55; 0.65; 0.9;
0.9)) 0.9)) 0.9)) 0.9))
C4 ((0.42; 0.62; 0.62; 0.82; 1;1), ((0.58; 0.76; 0.76; 0.88; 1;1), ((0.66; 0.84; 0.84; 0.94; 1;1), ((0.78; 0.94; 0.94; 1;1; 1), ((0.8; 0.95; 0.95; 1;1; 1), (0.88;
(0.52; 0.62; 0.62; 0.72; 0.9; 0.9)) (0.67; 0.76; 0.76; 0.82; 0.9; (0.75; 0.84; 0.84; 0.89; 0.9; (0.86; 0.94; 0.94; 0.97; 0.9; 0.95; 0.95; 0.98; 0.9; 0.9))
0.9)) 0.9)) 0.9))
C5 ((0.71; 0.9; 0.9; 1;1; 1), (0.81; ((0.51; 0.69; 0.69; 0.84; 1;1), ((0.63; 0.82; 0.82; 0.92; 1;1), ((0.35; 0.55; 0.55; 0.76; 1;1), ((0.59; 0.78; 0.78; 0.9; 1;1),
0.9; 0.9; 0.95; 0.9; 0.9)) (0.6; 0.69; 0.69; 0.77; 0.9; 0.9)) (0.72; 0.82; 0.82; 0.87; 0.9; (0.45; 0.55; 0.55; 0.65; 0.9; (0.68; 0.78; 0.78; 0.84; 0.9;
0.9)) 0.9)) 0.9))
C6 ((0.39; 0.59; 0.59; 0.78; 1;1), ((0.43; 0.61; 0.61; 0.76; 1;1), ((0.8; 0.94; 0.94; 1;1; 1), (0.87; ((0.29; 0.45; 0.45; 0.61; 1;1), ((0.55; 0.76; 0.76; 0.92; 1;1),
(0.49; 0.59; 0.59; 0.68; 0.9; 0.9)) (0.52; 0.61; 0.61; 0.68; 0.9; 0.94; 0.94; 0.97; 0.9; 0.9)) (0.37; 0.45; 0.45; 0.53; 0.9; (0.65; 0.76; 0.76; 0.84; 0.9;
0.9)) 0.9)) 0.9))
C7 ((0.52; 0.73; 0.73; 0.9; 1;1), ((0.42; 0.6; 0.6; 0.77; 1;1), ((0.67; 0.81; 0.81; 0.9; 1;1), ((0.6; 0.81; 0.81; 0.96; 1;1), ((0.69; 0.88; 0.88; 1;1; 1),
(0.63; 0.73; 0.73; 0.81; 0.9; 0.9)) (0.51; 0.6; 0.6; 0.69; 0.9; 0.9)) (0.74; 0.81; 0.81; 0.85; 0.9; (0.71; 0.81; 0.81; 0.89; 0.9; (0.78; 0.88; 0.88; 0.94; 0.9;
0.9)) 0.9)) 0.9))
C8 ((0.26; 0.36; 0.36; 0.6; 1;1), ((0.21; 0.25; 0.25; 0.33; 1;1), ((0.2; 0.24; 0.24; 0.33; 1;1), ((0.38; 0.56; 0.56; 1;1; 1) (0.45; ((0.24; 0.33; 0.33; 0.53; 1;1),
(0.31; 0.36; 0.36; 0.45; 0.9; 0.9)) (0.23; 0.25; 0.25; 0.29; 0.9; (0.22; 0.24; 0.24; 0.28; 0.9; 0.56; 0.56; 0.72; 0.9; 0.9)) (0.28; 0.33; 0.33; 0.41; 0.9;
0.9)) 0.9)) 0.9))
C9 ((0.4; 0.45; 0.45; 0.58; 1;1), ((0.51; 0.66; 0.66; 1;1; 1), ((0.41; 0.49; 0.49; 0.66; 1;1), ((0.41; 0.46; 0.46; 0.58; 1;1) ((0.4; 0.46; 0.46; 0.61; 1;1),
(0.43; 0.45; 0.45; 0.51; 0.9; 0.9)) (0.58; 0.66; 0.66; 0.79; 0.9; (0.45; 0.49; 0.49; 0.56; 0.9; (0.44; 0.46; 0.46; 0.51; 0.9; (0.43; 0.46; 0.46; 0.53; 0.9;
0.9)) 0.9)) 0.9)) 0.9))
C10 ((0.3; 0.5; 0.5; 0.7; 1;1), (0.4;0.5; ((0.22; 0.38; 0.38; 0.56; 1;1), ((0.82; 0.96; 0.96; 1;1; 1), ((0.38; 0.58; 0.58; 0.76; 1;1), ((0.5; 0.7; 0.7; 0.86; 1;1), (0.6;
0.5; 0.6; 0.9; 0.9)) (0.3; 0.38; 0.38; 0.47; 0.9; 0.9)) (0.89; 0.96; 0.96; 0.98; 0.9; (0.48; 0.58; 0.58; 0.67; 0.9; 0.7; 0.7; 0.78; 0.9; 0.9))
0.9)) 0.9))
C11 ((0.21; 0.26; 0.26; 0.36; 1;1), ((0.35; 0.53; 0.53; 1;1; 1), ((0.19; 0.22; 0.22; 0.29; 1;1), ((0.19; 0.21; 0.21; 0.27; 1;1), ((0.2; 0.23; 0.23; 0.31; 1;1),
(0.23; 0.26; 0.26; 0.3; 0.9; 0.9)) (0.42; 0.53; 0.53; 0.69; 0.9; (0.2; 0.22; 0.22; 0.25; 0.9; 0.9)) (0.2; 0.21; 0.21; 0.24; 0.9; 0.9)) (0.21; 0.23; 0.23; 0.26; 0.9;
0.9)) 0.9))

Table 10
The fuzzy weighted normalized decision matrix.

Criteria A1 Airport A2 Airport A3 Airport A4 Airport A5 Airport

C1 ((0.39; 0.65; 0.65; 0.86; 1;1), ((0.2; 0.42; 0.42; 0.65; 1;1), ((0.46; 0.71; 0.71; 0.88; 1;1), ((0.2; 0.4; 0.4; 0.61; 1;1) (0.29; ((0.37; 0.64; 0.64; 0.86; 1;1),
(0.51; 0.65; 0.65; 0.75; 0.9; (0.3;0.42; 0.42; 0.53; 0.9; 0.9)) (0.58; 0.71; 0.71; 0.79; 0.9; 0.4; 0.4; 0.5; 0.9; 0.9)) (0.49; 0.64; 0.64; 0.74; 0.9;
0.9)) 0.9)) 0.9))
C2 ((0.32; 0.59; 0.59; 0.86; 1;1), ((0.2; 0.43; 0.43; 0.71; 1;1), ((0.35; 0.61; 0.61; 0.84; 1;1), ((0.12; 0.24; 0.24; 0.42; 1;1), ((0.13; 0.34; 0.34; 0.59; 1;1),
(0.45; 0.59; 0.59; 0.72; 0.9; (0.31; 0.43; 0.43; 0.56; 0.9; (0.47; 0.61; 0.61; 0.72; 0.9; (0.18; 0.24; 0.24; 0.32; 0.9; (0.22; 0.34; 0.34; 0.46; 0.9;
0.9)) 0.9)) 0.9)) 0.9)) 0.9))
C3 ((0.56; 0.83; 0.83; 0.98; 1;1), ((0.21; 0.4; 0.4; 0.6; 1;1), ((0.5; 0.77; 0.77; 0.96; 1;1), ((0.19; 0.3; 0.3; 0.43; 1;1), ((0.26; 0.5; 0.5; 0.74; 1;1),
(0.69; 0.83; 0.83; 0.9; 0.9; 0.9)) (0.3;0.4; 0.4; 0.5; 0.9; 0.9)) (0.63; 0.77; 0.77; 0.86; 0.9; (0.24; 0.3; 0.3; 0.36; 0.9; 0.9)) (0.37; 0.5; 0.5; 0.61; 0.9; 0.9))
0.9))
C4 ((0.26; 0.51; 0.51; 0.79; 1;1), ((0.36; 0.62; 0.62; 0.84; 1;1), ((0.41; 0.69; 0.69; 0.9; 1;1), ((0.48; 0.77; 0.77; 0.96; 1;1), ((0.5; 0.78; 0.78; 0.96; 1;1),
(0.37; 0.51; 0.51; 0.64; 0.9; (0.48; 0.62; 0.62; 0.73; 0.9; (0.54; 0.69; 0.69; 0.79; 0.9; (0.62; 0.77; 0.77; 0.86; 0.9; (0.63; 0.78; 0.78; 0.87; 0.9;
0.9)) 0.9)) 0.9)) 0.9)) 0.9))
C5 ((0.26; 0.48; 0.48; 0.72; 1;1), ((0.18; 0.37; 0.37; 0.6; 1;1), ((0.23; 0.44; 0.44; 0.66; 1;1), ((0.12; 0.3; 0.3; 0.54; 1;1), ((0.21; 0.42; 0.42; 0.65; 1;1),
(0.36; 0.48; 0.48; 0.6; 0.9; 0.9)) (0.27; 0.37; 0.37; 0.48; 0.9; (0.33; 0.44; 0.44; 0.55; 0.9; (0.2;0.3; 0.3; 0.41; 0.9; 0.9)) (0.31; 0.42; 0.42; 0.53; 0.9;
0.9)) 0.9)) 0.9))
C6 ((0.22; 0.45; 0.45; 0.7; 1;1), ((0.25; 0.47; 0.47; 0.68; 1;1), ((0.46; 0.71; 0.71; 0.9; 1;1), ((0.17; 0.34; 0.34; 0.55; 1;1), ((0.32; 0.57; 0.57; 0.83; 1;1),
(0.33; 0.45; 0.45; 0.57; 0.9; (0.35; 0.47; 0.47; 0.57; 0.9; (0.58; 0.71; 0.71; 0.8; 0.9; 0.9)) (0.25; 0.34; 0.34; 0.44; 0.9; (0.44; 0.57; 0.57; 0.69; 0.9;
0.9)) 0.9)) 0.9)) 0.9))
C7 ((0.39; 0.66; 0.66; 0.88; 1;1), ((0.31; 0.54; 0.54; 0.76; 1;1), ((0.49; 0.73; 0.73; 0.88; 1;1), ((0.45; 0.73; 0.73; 0.94; 1;1), ((0.51; 0.79; 0.79; 0.98; 1;1),
(0.51; 0.66; 0.66; 0.76; 0.9; (0.42; 0.54; 0.54; 0.65; 0.9; (0.61; 0.73; 0.73; 0.8; 0.9; 0.9)) (0.58; 0.73; 0.73; 0.83; 0.9; (0.64; 0.79; 0.79; 0.88; 0.9;
0.9)) 0.9)) 0.9)) 0.9))
C8 ((0.17; 0.3; 0.3; 0.58; 1;1), ((0.14; 0.21; 0.21; 0.32; 1;1), ((0.13; 0.2; 0.2; 0.32; 1;1), ((0.25; 0.47; 0.47; 0.96; 1;1), ((0.16; 0.28; 0.28; 0.51; 1;1),
(0.23; 0.3; 0.3; 0.41; 0.9; 0.9)) (0.17; 0.21; 0.21; 0.26; 0.9; (0.16; 0.2; 0.2; 0.25; 0.9; 0.9)) (0.34; 0.47; 0.47; 0.65; 0.9; (0.21; 0.28; 0.28; 0.37; 0.9;
0.9)) 0.9)) 0.9))
C9 ((0.32; 0.43; 0.43; 0.58; 1;1), ((0.4; 0.62; 0.62; 1;1; 1), ((0.32; 0.46; 0.46; 0.66; 1;1), ((0.32; 0.44; 0.44; 0.58; 1;1), ((0.31; 0.44; 0.44; 0.61; 1;1),
(0.37; 0.43; 0.43; 0.49; 0.9; (0.5;0.62; 0.62; 0.77; 0.9; 0.9)) (0.38; 0.46; 0.46; 0.54; 0.9; (0.38; 0.44; 0.44; 0.5; 0.9; 0.9)) (0.37; 0.44; 0.44; 0.51; 0.9;
0.9)) 0.9)) 0.9))
C10 ((0.17; 0.38; 0.38; 0.63; 1;1), ((0.13; 0.29; 0.29; 0.5; 1;1), ((0.48; 0.73; 0.73; 0.9; 1;1), ((0.22; 0.44; 0.44; 0.68; 1;1), ((0.29; 0.53; 0.53; 0.77; 1;1),
(0.27; 0.38; 0.38; 0.5; 0.9; 0.9)) (0.2;0.29; 0.29; 0.39; 0.9; 0.9)) (0.6;0.73; 0.73; 0.81; 0.9; 0.9)) (0.32; 0.44; 0.44; 0.56; 0.9; (0.4;0.53; 0.53; 0.65; 0.9; 0.9))
0.9))
C11 ((0.11; 0.18; 0.18; 0.31; 1;1), ((0.17; 0.37; 0.37; 0.86; 1;1), ((0.09; 0.15; 0.15; 0.25; 1;1), ((0.09; 0.15; 0.15; 0.23; 1;1), ((0.1; 0.16; 0.16; 0.27; 1;1),
(0.14; 0.18; 0.18; 0.23; 0.9; (0.25; 0.37; 0.37; 0.54; 0.9; (0.12; 0.15; 0.15; 0.2; 0.9; 0.9)) (0.12; 0.15; 0.15; 0.19; 0.9; (0.13; 0.16; 0.16; 0.21; 0.9;
0.9)) 0.9)) 0.9)) 0.9))
M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99 93

!
aU aU aU aU  U  U
~rij ¼ i1 i2 i3 i4 ~
; ; ; ; H1 Ai ; H2 Ai ~ ;
2 cj cj cj cj
14X3 !!
~1 ¼
w ð0:70; 0:90; 0:90; 1:0; 1:0; 1:0Þ aLi1 aLi2 aLi3 aLi4  L  L
5 y¼1 ~
; ; ; ; H1 A1i ; H2 Ai ~
cj cj cj cj
þð0:50; 0:70; 0:70; 0:90; 1:0; 1:0Þ   
4:20 6:20 6:20 8:20 5:20 6:20 6:20 7:20
# ¼ ; ; ; ; 1; 1 ; ; ; ; ; 0:9; 0:9
10 10 10 10 10 10 10 10
þð0:90; 1:0; 1:0; 1:0; 1:0; 1:0Þ þ … ;
¥ð0:42; 0:62; 0:62; 0:82; 1; 1Þ; ð0:52; 0:62; 0:62; 0:72; 0:9; 0:9Þ
2
14X3
The fuzzy normalized value of alternative A1 for criterion C9
ð0:80; 0:90; 0:9; 0:95; 0:9; 0:9Þ (Distance and costs) is given by (cost criteria)
5 y¼1

þð0:60; 0:70; 0:70; 0:80; 0:9; 0:9Þ c


j ¼ min ð3:80; 3:80; 3:80; 3:80Þ
# i
þð0:95; 1:0; 1:0; 1:0; 0:9; 0:9Þ þ … !
c     U  U
j cj cj cj ~ ;H A ~
~rij ¼ ; ; ; ; H 1 A i 2 i ;
¼ ð0:58; 0:76; 0:76; 0:88; 1:0; 1:0Þ; aU U
i4 ai3 ai2 ai1
U U

!!
ð0:67; 0:76; 0:76; 0:82; 0:9; 0:9Þ c     L  L
j cj cj cj ~ ;H A ~
; ; ; ; H 1 A 1i 2 i
The fuzzy decision matrix for the alternatives is computed by aLi4 aLi3 aLi2 aLi1
using Eqs. (2)e(4) in Table 8. Next step is computed the fuzzy   
3:80 3:80 3:80 3:80 3:80 3:80 3:80 3:80
normalized decision matrix of alternatives using Eqs. (5)e(7). The ¼ ; ; ; ; 1; 1 ; ; ; ; ; 0:9; 0:9
6:60 8:40 8:40 9:40 7:50 8:40 8:40 8:90
fuzzy normalized decision matrix for the five alternatives is pre-
sented in Table 9. The fuzzy normalized value of alternative A1 for ¥ð0:40; 0:45; 0:45; 0:58; 1; 1Þ; ð0:43; 0:45; 0:45; 0:51; 0:9; 0:9Þ
criterion C4 (Number of transit passengers) is given by (benefit
c ¼ max ð10:0; 10:0; 10:0; 10:0Þ Similarly, the normalized values of the alternatives are
criteria) j computed for the remaining criteria. The later weight of each cri-
i
terion weighted normalized fuzzy decision matrix is found in
Table 11 Table 10. This matrix is computed with Eq. (8). For alternative A1,
~ of the trapezoidal interval Type-2 fuzzy set.
The ranking value Rank A i
the fuzzy weight for criterion C11 (Seasonality) is given by
Criteria A1 A2 A3 A4 A5 PIS NIS ~ij ¼ ~rij ð:Þw
v ~j
C1 7.503 6.160 7.868 6.077 7.407 7.868 6.077
C2 7.160 6.246 7.252 5.178 5.689 7.252 5.178 Table 13
C3 8.524 6.089 8.201 5.520 6.608 8.524 5.520 Closeness coefficients (CCj) of the five alternatives.
C4 6.685 7.335 7.707 8.182 8.233 8.233 6.685
C5 6.550 5.925 6.295 5.488 6.174 6.550 5.488 Alternatives D* D- CCj Ranking
C6 6.347 6.436 7.877 5.735 7.061 7.877 5.735 A1 Airport 9.529 9.995 0.512 2
C7 7.526 6.886 7.982 7.955 8.293 8.293 6.886 A2 Airport 13.575 5.949 0.305 4
C8 5.569 5.022 4.989 6.552 5.428 6.552 4.989 A3 Airport 5.176 14.348 0.735 1
C9 6.288 7.396 6.472 6.343 6.341 7.396 6.288 A4 Airport 14.483 5.041 0.258 5
C10 5.949 5.438 7.965 6.296 6.819 7.965 5.438 A5 Airport 9.688 9.836 0.504 3
C11 4.848 5.970 4.692 4.670 4.739 5.970 4.670

Table 12
Distances d(Aj, A*) and d(Aj, A-) of the alternatives from fuzzy positive ideal solution (FPIS) and fuzzy negative ideal solution (FNIS) (i,j ¼ 1,2,3,4,5).

Criteria FPIS FNIS

A1 A2 A3 A4 A5 A1 A2 A3 A4 A5

C1 0.365 1.708 0.000 1.792 0.462 1.426 0.084 1.792 0.000 1.330
C2 0.091 1.005 0.000 2.074 1.563 1.983 1.069 2.074 0.000 0.511
C3 0.000 2.435 0.322 3.003 1.916 3.003 0.569 2.681 0.000 1.087
C4 1.548 0.898 0.525 0.051 0.000 0.000 0.650 1.023 1.497 1.548
C5 0.000 0.625 0.254 1.062 0.376 1.062 0.437 0.807 0.000 0.686
C6 1.530 1.441 0.000 2.142 0.816 0.612 0.701 2.142 0.000 1.326
C7 0.767 1.407 0.311 0.338 0.000 0.640 0.000 1.097 1.069 1.407
C8 0.982 1.530 1.563 0.000 1.123 0.580 0.033 0.000 1.563 0.439
C9 1.108 0.000 0.924 1.053 1.055 0.000 1.108 0.184 0.055 0.052
C10 2.016 2.527 0.000 1.669 1.146 0.511 0.000 2.527 0.858 1.381
C11 1.122 0.000 1.277 1.299 1.231 0.177 1.299 0.022 0.000 0.068
94 M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99

Fig. 5. Ranking the evaluated alternatives.

Similarly, the fuzzy weights of the five alternatives for the


remaining criteria are calculated. The positive and the negative-
ideal solutions are computed by using Eqs. (9) and (10). Table 11
shows the results of this step. Then, the Euclidean distance of
each alternative from A* and A can be computed using Eqs. (11)
and (12), as shown in Table 12. Subsequently, the closeness coef-
ficient of each alternative is computed by Eq. (13). For example, the
calculation of closeness coefficient for the alternative A1 is as
follows:

d
1 9:995
CC1 ¼ ¼ ¥0:512
d1 þ d
1 9:529 þ 9:995
Finally, the values of each alternative for final ranking are shown
in Table 13.
By solving the decision problem for new routes via interval
type-2 fuzzy TOPSIS method, the resulting alternatives are ranked
Fig. 6. Evaluation of D* & D for alternatives.
and shown in Fig. 5. According to final ranking of A3 > A1> A5> A2>
A4. A3 is the best among the five alternatives because it has the
largest closeness coefficient (CCj), while A4 is the worst
¼ ð0:21; 0:26; 0:26; 0:36; 1; 1; 0:23; 0:26; 0:26; 0:30; 0:9; 0:9Þ alternative.
ð:Þð0:50; 0:70; 0:70; 0:86; 1; 1; 0:60; 0:70; 0:70; 0:78; 0:9; 0:9Þ Comparison of D*1, D*2 …, D*5 and D-1, D-2 …, D-5 respectively is
y ð0:11; 0:18; 0:18; 0:31; 1; 1; 0:14; 0:18; 0:18; 0:23; 0:9; 0:9Þ: shown in Fig. 6. For example, it can be seen that A3 has large D i
which shows large distance from negative ideal.

Table 14
The CCj value of each alternative by using each resolving coefficient value.

Experiment number CCj Description

A1 A2 A3 A4 A5

1 0.517 0.405 0.605 0.259 0.487 All criteria weights ¼ ((0, 0, 0, 0.1; 1, 1), (0, 0, 0, 0.05; 0.9, 0.9))
2 0.525 0.333 0.717 0.246 0.506 All criteria weights ¼ ((0, 0.1, 0.1, 0.3; 1, 1), (0.05, 0.1, 0.1, 0.2; 0.9, 0.9))
3 0.524 0.321 0.732 0.244 0.506 All criteria weights ¼ ((0.1, 0.3, 0.3, 0.5; 1, 1), (0.2, 0.3, 0.3, 0.4; 0.9, 0.9))
4 0.523 0.316 0.737 0.243 0.505 All criteria weights ¼ ((0.3, 0.5, 0.5, 0.7; 1, 1), (0.4, 0.5, 0.5, 0.6; 0.9, 0.9))
5 0.522 0.314 0.739 0.242 0.505 All criteria weights ¼ ((0.5, 0.7, 0.7, 0.9; 1, 1), (0.6, 0.7, 0.7, 0.8; 0.9, 0.9))
6 0.522 0.313 0.742 0.242 0.505 All criteria weights ¼ ((0.7, 0.9, 0.9, 1; 1, 1), (0.8, 0.9, 0.9, 0.95; 0.9, 0.9))
7 0.522 0.311 0.744 0.241 0.505 All criteria weights ¼ ((0.9, 1, 1, 1; 1, 1), (0.95, 1, 1, 1; 0.9, 0.9))
8 0.775 0.069 0.974 0.017 0.721 Weight of criteria 1¼ ((0.9, 1, 1, 1; 1, 1), (0.95, 1, 1, 1; 0.9, 0.9))
Weight of remaining criteria ¼ ((0, 0, 0, 0.1; 1, 1), (0, 0, 0, 0.05; 0.9, 0.9))
9 0.929 0.504 0.978 0.015 0.254 Weight of criteria 2 ¼ ((0.9, 1, 1, 1; 1, 1), (0.95, 1, 1, 1; 0.9, 0.9))
Weight of remaining criteria ¼ ((0, 0, 0, 0.1; 1, 1), (0, 0, 0, 0.05; 0.9, 0.9))
10 0.296 0.330 0.978 0.015 0.608 Weight of criteria 6 ¼ ((0.9, 1, 1, 1; 1, 1), (0.95, 1, 1, 1; 0.9, 0.9))
Weight of remaining criteria ¼ ((0, 0, 0, 0.1; 1, 1), (0, 0, 0, 0.05; 0.9, 0.9))
11 0.215 0.020 0.981 0.333 0.541 Weight of criteria 10¼ ((0.9, 1, 1, 1; 1, 1), (0.95, 1, 1, 1; 0.9, 0.9))
Weight of remaining criteria ¼ ((0, 0, 0, 0.1; 1, 1), (0, 0, 0, 0.05; 0.9, 0.9))
M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99 95

Fig. 7. Sensitivity analysis of CCj values for each alternative.

4.1.1. Sensitivity analysis enabling both the airport and airline to succeed in a long-term
In the study, a sensitivity analysis has been conducted to show partnership. As can be seen above (see Table 7), C9 and C3 are
changes in case different priorities regarding the choice of new important criteria to determine in order to open a new route. Ac-
route are altered. For this aim, the very high value or the very low cording to them, A3 (SFO) is chosen as an optimal and efficient new
value of each criteria has been altered and 11 experimental com- destination. They will affect the region where A3 is located. There
binations have been formed and for each case the CCj values have will be an increase in the total passengers to/from A3. It means that
been re-calculated. The results of the sensitivity analysis can be increased awareness of the new route (A3) will be widespread
seen in Table 14, while the graphical display of these results is as around the world. The airline company can benefit from the
shown in Fig. 7. As a result of the eleven experiments conducted, A3 outcome and it will add value to their current profit. In other words,
once again came out as the best alternative. The alternatives of A1, there will be revenue increases. In this study, five expert opinions
A2 and A5 went through substantial changes in experiment 8, 9 and from different departments have been taken into account in order
10. This shows that the criterion priorities of these alternatives are to find out the important criteria. We have applied fuzzy-based
sensitive in these experiments. MCDM approaches, including interval type-2 Fuzzy TOPSIS and
type-I Fuzzy TOPSIS, comparing their performance to decide the
best destination in the North America region, which has not been
5. Results and discussion studied before. In fact, this problem can be solved by using any of
these two methods, but given the importance of selection of new
In this study, we aim to find the best route which is sustainable,

Fig. 8. Result of proposed methodologies.


96 M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99

route problem, the best alternative is searched by testing many 6. Conclusion


techniques. Furthermore, the elasticity of these methods is also
compared to each other. One of the most important factors that affect the profitability in
The new route selection problem can be addressed by applying airline companies is the new route selection and capacity planning
the type-1 fuzzy TOPSIS method in order to compare the results, of current routes. The new route selection requires making the
shown in Fig. 8. The ranking of each of the different routes obtained correct choice on the first attempt because of its irreversible effect
from interval type-2 fuzzy TOPSIS is A3 > A1, and A3 > A1 is also in the long term. Even it can be reversed, the wrong decision can
obtained by the type-1 fuzzy TOPSIS method. These are similar cause huge costs, especially in long haul markets. In route analysis
rankings despite that they are based on different fuzzy sets and problems, where the route choice involves multiple factors that
solution steps. This method is proposed as a useful way to handle interact with each other and most of the time the desired gains are
fuzzy multi criteria group decision-making problems in a more contradicting, the choice of methods leading to a conciliating so-
flexible and smarter manner since it uses interval type-2 fuzzy sets lution is ever more important. For this reason, the necessity of
rather than type-1 fuzzy sets to represent the evaluating values and methods affecting the final decision that are utilized in reaching a
the weights of attributes (Chen and Lee, 2010a). solution, is of vital importance in regards to reaching the effective
As a result, destination of San Francisco (A3) stands out as the solution.
most significant among other alternative destinations. There are This study focuses on a new route analysis of an airline com-
both advantages and disadvantages of this city, and among the pany operating in Turkey which plans opening in one of five air-
advantages the following can be listed: ports in the North America. In line with this, in the decision
making process of the new route, a decision maker group con-
 A3 is a high technology city and Stanford University, as one of the sisting of experts has been formed and the interval type-2 fuzzy
most eminent universities of the world, is also located in this TOPSIS method has been used to solve this problem. This method
city. has been implemented on a real-world problem. Interval type-2
 It has a quite high number of Indian and Middle East passengers. fuzzy TOPSIS method detects A3 airport (SFO) as the best alterna-
 It has high revenue potential due to having high tech start-up tive for route in North America based on the specified set of
companies. criteria. The results show that interval type-2 fuzzy MCDM
approach is a useful tool for decision makers in the selection of
The disadvantages can be listed as the following: potential airports for route analysis. Since the method used in this
study is used for the first time in the literature, it can shed light on
 It has the highest cost among the alternative destinations due to future studies as well.
its distance from Istanbul. The destination is located at the most This study can be of help for those who want to examine new
western part of North American Region. markets, potential passengers and generating revenue for both new
 It has a high competition level as Emirates and Etihad also flies and current airlines and airports in Turkey and North America.
to this destination, which can carry passengers to similar re- However, the model has the following limitations; some of the
gions to A3. criteria, i.e yields, local and transfer demand, cost could not be
declared due to confidentiality. Also, non-demand related factors
Due to network structure of the airline company, North such as economic condition of potential country, political relations
American flights have connections with African, South East between countries, social factors, security and safety measure-
Asian and Middle East regions. According to Goedeking (2010), ments affect the decision of a new route.
passengers are expected to make transfers for long-haul trips Proposed MCDM model provides a constructed methodology for
having a detour factor of less than 1.2. Whereas long-haul flights demand related data and ranks the potential destinations in
have lower detour factors, short-haul trips have higher detour managerial perspective. Prioritization of the potential routes could
factors, rising up to 2 in some cases. For instance, the flight from be used for allocating the marketing, research, human resources of
San Francisco to Mumbai (SFO-BOM) directly flies to Mumbai. If the company. On the other hand, the model could be also modified
this flight indirectly goes to Mumbai via Istanbul, the detour and utilized for the aircraft selection for the existing and new
factor is calculated as 1.15, which makes sense according to routes, engine selection of the newly ordered aircrafts, codeshare or
Geodeking. joint venture partner selection. Also it will help to manage the
As can be seen in the above example, the competition levels in company's medium and long term growth plan. This methodology
regions having significant detour factor values are summarized in could be applied to other regions and can be beneficial to identify
Table 15 of the appendix. The table indicates that there are few potential routes for upcoming aircrafts as well.
flights to South East Asia, Africa and Turkey. In the Middle East
region, on the other hand, Emirates, Qatar and Etihad airline
companies have flights to certain destinations. Appendix
Weekly trip numbers between North America region and Turkey
are presented in Table 16 of the appendix. When the flights be-
tween North America and Turkey are examined, only in Toronto
location, there is a competition between Air Canada and Turkish
Table 5
Airlines. Furthermore, Delta Airlines provides a capacity which is
Linguistic assessments for the 11 criteria.
one third of what Turkish Airlines provides in New York location
only during peak summer months. Decision Criteria
makers
The comparison of weekly trips to North American region of C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11
Gulf Carriers and Turkish Airlines is summarized in Table 17 of the DM1 H VH H MH L MH H H H VH H
appendix. New York is a highly competitive destination point. In DM2 MH M VH H M M VH MH VH M M
Seattle destination, Emirates has a monopoly power with its fre- DM3 VH MH VH H M VH H VH H MH M
quency of 14 weekly flights. In other destinations, airline com- DM4 VH VH VH M MH H VH MH VH H MH
DM5 H M H MH H MH MH H VH H MH
panies generally serve 7 trips weekly.
M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99 97

Table 6
Linguistic assessments for the five alternatives.

Alternatives Decision Criteria


makers
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11

A1 Airport DM1 VG G G F VG G G G G F G
DM2 G G G MG G MG G F G F F
DM3 MG MG VG F MG F F MP VG MP MP
DM4 MG MG MG MG G MP F MP F F G
DM5 G F VG MG G F MG F G MG G

A2 Airport DM1 MG MG F F G G MG F G F MP
DM2 F F VP VG MG MP G VG MP MP MP
DM3 MP MP F G VG F G G F P P
DM4 G MG G G MP VG P MG G G G
DM5 MP F MP F F MP MP F MP P P

A3 Airport DM1 VG VG MG G G VG VG G F VG MG
DM2 VG G G F G MG G MG G VG G
DM3 G G VG G VG G G MG G G G
DM4 MG MP MG G MP VG P F G G G
DM5 VG MG G VG G VG VG G MG VG MG

A4 Airport DM1 MP P VP G MG VP MG MP G MP G
DM2 F VP MP G F MP G P VG F VG
DM3 F MP VP VG F F G VP F F MG
DM4 VG VG VG G MG G G G VG G G
DM5 MP P VP VG MP F F MP MG MG MG

A5 Airport DM1 G F F VG G G G F G G G
DM2 G F MP VG VG MG VG MG G MG G
DM3 MG MP F G MG F MG F MG F F
DM4 G MP MG G MP G MG MP MG G G
DM5 MG F MG VG G MG G MG G F MG

Table 16
Weekly frequency between North American region and Turkey as of 23 September
2015 (http://centreforaviation.com).

City Weekly frequency Notes


Table 15
New York JFK 21 Monopoly
Competition in alternative destinations and regions having lower detour factors via
San Francisco 7 Monopoly
Istanbul.
Los Angeles 7 Monopoly
Alternative Regions Airline companies Chicago O'Hare 7 Monopoly
Houston G. Bush 7 Monopoly
TR AF ME SA Emirates Qatar Etihad Toronto Pearson 6 Air Canada is only competitor
DEN x x x x x x x Washington Dulles 7 Monopoly
MIA þ x þ x x þ x Boston Logan 7 Monopoly
SFO þ x þ þ þ x þ Montreal Trudeau 3 Monopoly
BOS þ x þ x þ þ x Future plans
DFW x x þ x þ þ þ Miami 7 Monopoly
Atlanta 7 Monopoly
98 M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99

Table 17
Comparison of weekly trips to North American region of Gulf Carriers and Turkish Airlines (http://centreforaviation.com).

THY from IST Emirates from Dubai Qatar airways from Doha Etihad from Abu Dhabi

New York JFK 21 21 7 14


San Francisco 7 7 e 7
Los Angeles 7 7 e 7
Chicago O'Hare 7 7 7 7
Houston G. Bush 7 7 7 e
Toronto Pearson 6 3 e 3
Washington Dulles 7 7 7 7
Boston Logan 7 7 e e
Montreal Trudeau 3 e 3 e
Seattle/Tacoma e 14 e e
Dallas/Fort Worth e 7 7 3
Orlando e 7 e
Philadelphia e e 7 e
Miami 7 e 5 e
Atlanta 7 e 7

References Goedeking, P., 2010. Networks in Aviation: Strategies and Structures. Springer Sci-
ence & Business Media.
Gong, Y., 2013. Fuzzy multi-attribute group decision making method based on in-
Air Travel Demand, 2008. IATA ECONOMICS BRIEFING, No:9. http://www.iata.org/
terval type-2 fuzzy sets and applications to global supplier selection. Int. J.
publications/economic-briefings/air_travel_demand.pdf.
Fuzzy Syst. 15 (4), 392.
Ashtiani, B., Haghighirad, F., Makui, A., Ali Montazer, G., 2009. Extension of fuzzy
Hu, J., Zhang, Y., Chen, X., Liu, Y., 2013. Multi-criteria decision making method based
TOPSIS method based on interval-valued fuzzy sets. Appl. Soft Comput. 9 (2),
on possibility degree of interval type-2 fuzzy number. Knowledge-Based Syst.
457e461.
43, 21e29.
Barros, C.P., Wanke, P., 2015. An analysis of African airlines efficiency with two-stage
Hwang, C.L., Yoon, K., 1981. Multiple Attributes Decision Making Methods and
TOPSIS and neural networks. J. Air Transp. Manag. 44, 90e102.
Applications. Springer, Berlin Heidelberg.
Bratu, S., 1998. Network Value Concept in Airline Revenue Management. Massa-
IADS (International Aviation Developments Series), 2014. U.S. International Pas-
chusetts Institute of Technology], Flight Transportation Laboratory, [Cambridge,
senger and Freight Statistics. https://www.transportation.gov/sites/dot.gov/
Mass.
files/docs/US%20International%20Air%20Passenger%20and%20Freight%
Brueckner, J.K., Spiller, P.T., 1994. Economies of traffic density in the deregulated
20Statistics%20Report%20for%20December%202014_0.pdf.
airline industry. J. Law Econ. 379e415.
Jaillet, P., Song, G., Yu, G., 1996. Airline network design and hub location problems.
Caves, D.W., Christensen, L.R., Tretheway, M.W., 1984. Economies of density versus
Locat. Sci. 4 (3), 195e212.
economies of scale: why trunk and local service airline costs differ. RAND J. € _
Kahraman, C., Oztayşi, B., Sarı, I.U., lu, E., 2014. Fuzzy analytic hierarchy
Turanog
Econ. 471e489.
process with interval type-2 fuzzy sets. Knowledge-Based Syst. 59, 48e57.
Celik, E., Bilisik, O.N., Erdogan, M., Gumus, A.T., Baracli, H., 2013. An integrated novel
Kilic, M., Kaya, I., 2015. Investment project evaluation by a decision making
interval type-2 fuzzy MCDM method to improve customer satisfaction in public
methodology based on type-2 fuzzy sets. Appl. Soft Comput. 27, 399e410.
transportation for Istanbul. Transp. Res. Part E Logist. Transp. Rev. 58, 28e51.
Kuo, M.S., 2011. A novel interval-valued fuzzy MCDM method for improving airlines'
Celik, E., Aydin, N., Gumus, A.T., 2014. A multiattribute customer satisfaction eval-
service quality in Chinese cross-strait airlines. Transp. Res. Part E: Logist. Transp.
uation approach for rail transit network: a real case study for Istanbul, Turkey.
Rev. 47 (6), 1177e1193.
Transp. Policy 36, 283e293.
Lee, L.W., Chen, S.M., 2008. Fuzzy multiple attributes group decision-making based
Cevik Onar, S., Oztaysi, B., Kahraman, C., 2014. Strategic decision selection using
on the extension of TOPSIS method and interval type-2 fuzzy sets. In: Machine
hesitant fuzzy TOPSIS and interval type-2 fuzzy AHP: a case study. Int. J.
Learning and Cybernetics, 2008 International Conference on, vol. 6. IEEE,
Comput. Intell. Syst. 7 (5), 1002e1021.
pp. 3260e3265.
Çiftçi, M.E., Şevkli, M., 2015. A new hub and spoke system proposal: a case study for
Liao, T.W., 2015. Two interval type 2 fuzzy TOPSIS material selection methods.
Turkey's aviation industry. J. Air Transp. Manag. 47, 190e198.
Mater. Des. 88, 1088e1099.
Chang, Y.H., Yeh, C.H., 2002. A survey analysis of service quality for domestic air-
Liou, J.J., Chuang, Y.T., 2010. Developing a hybrid multi-criteria model for selection
lines. Eur. J. Oper. Res. 139 (1), 166e177.
of outsourcing providers. Expert Syst. Appl. 37 (5), 3755e3761.
Chen, S.M., Lee, L.W., 2010a. Fuzzy multiple attributes group decision-making based
Liou, J.J., Tzeng, G.H., Tsai, C.Y., Hsu, C.C., 2011. A hybrid ANP model in fuzzy envi-
on the interval type-2 TOPSIS method. Expert Syst. Appl. 37 (4), 2790e2798.
ronments for strategic alliance partner selection in the airline industry. Appl.
Chen, S.M., Lee, L.W., 2010b. Fuzzy multiple attributes group decision-making based
Soft Comput. 11 (4), 3515e3524.
on the ranking values and the arithmetic operations of interval type-2 fuzzy
Liou, J.J., 2012. Developing an integrated model for the selection of strategic alliance
sets. Expert Syst. Appl. 37 (1), 824e833.
partners in the airline industry. Knowledge-Based Syst. 28, 59e67.
Chen, Y.H., Tseng, M.L., Lin, R.J., 2011. Evaluating the customer perceptions on in-
Mardani, A., Jusoh, A., MD Nor, K., Khalifah, Z., Zakwan, N., Valipour, A., 2015a.
flight service quality. Afr. J. Bus. Manag. 5 (7), 2854.
Multiple criteria decision-making techniques and their applicationsea review
Chen, C.T., 2000. Extensions of the TOPSIS for group decision-making under fuzzy
of the literature from 2000 to 2014. Econ. Research-Ekonomska Istra zivanja 28
environment. Fuzzy Sets Syst. 114 (1), 1e9.
(1), 516e571.
Chen, T.Y., 2012. Comparative analysis of SAW and TOPSIS based on interval-valued
Mardani, A., Jusoh, A., Zavadskas, E.K., 2015b. Fuzzy multiple criteria decision-
fuzzy sets: discussions on score functions and weight constraints. Expert Syst.
making techniques and applicationseTwo decades review from 1994 to 2014.
Appl. 39 (2), 1848e1861.
Expert Syst. Appl. 42 (8), 4126e4148.
Demirel, N.Ç., Deveci, M., Eser, G., 2016. Comparative analysis of fuzzy multi-criteria
Mardani, A., Zavadskas, E.K., Khalifah, Z., Jusoh, A., Nor, K.M., 2015c. Multiple criteria
decision making for location selection of textile plant in Turkey. In: Proceedings
decision-making techniques in transportation systems: a systematic review of
of 24th International Academic Conferences (No. 4006524). International
the state of the art literature. Transport 1e27.
Institute of Social and Economic Sciences (IISES), Barcelona, ISBN 978-80-
Mendel, J.M., John, R., Liu, F., 2006. Interval type-2 fuzzy logic systems made simple.
87927-25-0. http://dx.doi.org/10.20472/IAC.2016.024.017.
€ Fuzzy Systems. IEEE Trans. 14 (6), 808e821.
Deveci, M., Demirel, N.Ç., John, R., Ozcan, E., 2015. Fuzzy multi-criteria decision
Mokhtarian, M.N., Sadi-nezhad, S., Makui, A., 2014. A new flexible and reliable in-
making for carbon dioxide geological storage in Turkey. J. Nat. Gas Sci. Eng. 27,
terval valued fuzzy VIKOR method based on uncertainty risk reduction in de-
692e705.
€ _ cision making process: an application for determining a suitable location for
DPT, 2001. Ulaştırma Ozel Ihtisas Komisyonu Raporu Hava Yolu Ulaştırması Alt
digging some pits for municipal wet waste landfill. Comput. Ind. Eng. 78,
Komisyonu Raporu, Sekizinci Beş Yıllık Kalkınma Planı (Ankara).
213e233.
Dymova, L., Sevastjanov, P., Tikhonenko, A., 2015. An interval type-2 fuzzy extension
Prakash, C., Barua, M.K., 2015. Integration of AHP-TOPSIS method for prioritizing the
of the TOPSIS method using alpha cuts. Knowledge-Based Syst. 83, 116e127.
_ solutions of reverse logistics adoption to overcome its barriers under fuzzy
Ertürk, H., 2011. Başa Baş Analizinin Hava Yolu Endüstrisi Için Uygulanması Ve Bir
€ _
i Sonuçların Irdelenmesi. environment. J. Manuf. Syst. 37, 599e615.
Ornek Hava Yolu Şirketi Üzerinde Bu Analizin Verdig
Prakash, C., Barua, M.K., 2016a. An analysis of integrated robust hybrid model for
Havayolu Ekonomisi.
third-party reverse logistics partner selection under fuzzy environment. Resour.
Garg, C.P., 2016. A robust hybrid decision model for evaluation and selection of the
Conserv. Recycl. 108, 63e81.
strategic alliance partner in the airline industry. J. Air Transp. Manag. 52, 55e66.
Prakash, C., Barua, M.K., 2016b. A combined MCDM approach for evaluation and
Ghorabaee, M.K., 2016. Developing an MCDM method for robot selection with in-
selection of third-party reverse logistics partner for Indian electronics industry.
terval type-2 fuzzy sets. Robot. Computer-Integrated Manuf. 37, 221e232.
M. Deveci et al. / Journal of Air Transport Management 59 (2017) 83e99 99

Sustain. Prod. Consum. 7, 66e78. Wanke, P., Barros, C.P., Chen, Z., 2015. An analysis of Asian airlines efficiency with
Prakash, C., Barua, M.K., 2016. A robust multi-criteria decision-making framework two-stage TOPSIS and MCMC generalized linear mixed models. Int. J. Prod.
for evaluation of the airport service quality enablers for ranking the airports. Econ. 169, 110e126.
J. Qual. Assur. Hosp. Tour. 1e20. Yavuz, S., Deveci, M., 2014. Bulanık TOPSIS ve bulanık VIKOR yo €ntemleriyle alışveriş
Qin, J., Liu, X., Pedrycz, W., 2015. An extended VIKOR method based on prospect merkezi kuruluş yeri seçimi ve bir uygulama. Ege Akad. Bakış 14 (3), 463e479.
theory for multiple attribute decision making under interval type-2 fuzzy Yoon, K.P., Hwang, C.L., 1995. Multiple Attribute Decision Making: an Introduction,
environment. Knowledge-Based Syst. 86, 116e130. Vol. 104. Sage publications.
Tan, C., Chen, X., 2013. Interval-valued intuitionistic fuzzy multicriteria group de-
cision making based on VIKOR and Choquet integral. J. Appl. Math. 2013.
e
THY (Türk Hava Yolları A.O.), 2009. THY Faaliyet Raporu, Seri:XI No:29 Sayılı Teblig Further reading
_
Istinaden Hazırlanmış Yo€ netim Kurulu Faaliyet Raporu.
e
THY (Türk Hava Yolları A.O.), 2011. THY Faaliyet Raporu, Seri:XI No:29 Sayılı Teblig Aviationeconomics, http://www.aviationeconomics.com.
_
Istinaden Hazırlanmış Yo€ netim Kurulu Faaliyet Raporu, 2011. Centreforaviation, http://centreforaviation.com.
Wang, W., Liu, X., Qin, Y., 2012. Multi-attribute group decision making models Climatepolicyinfohub, http://climatepolicyinfohub.eu.
under interval type-2 fuzzy environment. Knowledge-Based Syst. 30, 121e128. Dailymail, http://www.dailymail.co.uk.

You might also like