Economic Research-Ekonomska Istraživanja

ISSN: 1331-677X (Print) 1848-9664 (Online) Journal homepage: http://www.tandfonline.com/loi/rero20

Multiple criteria decision-making KEMIRA-M

method for solution of location alternatives

Aleksandras Krylovas, Edmundas Kazimieras Zavadskas & Natalja Kosareva

To cite this article: Aleksandras Krylovas, Edmundas Kazimieras Zavadskas & Natalja Kosareva
(2016) Multiple criteria decision-making KEMIRA-M method for solution of location alternatives,
Economic Research-Ekonomska Istraživanja, 29:1, 50-65, DOI: 10.1080/1331677X.2016.1152560

To link to this article: http://dx.doi.org/10.1080/1331677X.2016.1152560

© 2016 The Author(s). Published by Taylor &

Francis

Published online: 13 Apr 2016.

Submit your article to this journal

Article views: 584

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=rero20

Download by: [95.168.107.53] Date: 28 December 2016, At: 06:38

Economic Research-Ekonomska Istraživanja, 2016
VOL. 29, NO. 1, 50–65
http://dx.doi.org/10.1080/1331677X.2016.1152560

OPEN ACCESS

Multiple criteria decision-making KEMIRA-M method for

solution of location alternatives
Aleksandras Krylovasa, Edmundas Kazimieras Zavadskasb and Natalja Kosarevaa
a
Department of Mathematical Modelling, Vilnius Gediminas Technical University, Vilnius, Lithuania; bResearch
Institute of Smart Building Technologies, Vilnius Gediminas Technical University, Vilnius, Lithuania

ABSTRACT ARTICLE HISTORY

Choice of location in many cases is a key factor setting up a new Received 21 September 2015
business object. In this article the KEMIRA-M method is proposed to Accepted 7 February 2016
establish priority of criteria and determine criteria weights. Weighted KEYWORDS
sum of criteria values was applied for ranking the alternatives. This KEMIRA-M; multiple
technique is useful if the evaluation criteria naturally consist of several criteria decision-making;
logically explained groups of criteria. Method requires much less initial optimisation; location
information and is based upon searching the solution of optimisation alternatives
problem. KEMIRA-M is applied for the case study of construction site
for non-hazardous waste incineration plant in Vilnius City. JEL CLASSIFICATION
Q53; C44; D70

1. Introduction
The problem of choosing the best alternative location among several alternatives is relevant
in many fields of human economic activity such as marketing, capital investments, logistics,
construction, cultural heritage management, hospitality management, location of facilities,
etc. Optimal solutions of the location problem can help not only by saving money and
resources, but also by improving the environmental situation. A number of articles devoted
to this problem have been published recently.
Van Asperen and Dekker (2013) evaluated a number of alternative strategies of port-
of-entry choices by means of simulation. The optimal solution results in the lowest cost
per container. Broll, Roldán-Ponce, and Wahl (2013) investigated the impact of economic
risk and risk preferences upon regional allocation of capital investments. Gilvear, Spray,
and Casas-Mulet (2013) presented methodology of optimising the outcomes of river reha-
bilitation in terms of delivery of multiple ecosystem services. In the article, trajectories
over time for attaining the long-term ecosystem service score for each river rehabilitation
measures are given. Ansar (2013) analysed the procurement of infrastructure services prior
to making durable and immobile investments by large firms through a case study of a
large manufacturing firm, ThyssenKrupp AG. Gerritse and Moreno-Monroy (2012) built
a modified Core-Periphery model where formal and informal firms compete in consumer

CONTACT Edmundas Kazimieras Zavadskas [email protected]

© 2016 The Author(s). Published by Taylor & Francis.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/
licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
 Economic Research-Ekonomska Istraživanja  51

markets. Fuzzy Data Envelopment Analysis was used for measuring the relative efficiency
of foreign direct investment in 12 transition economies that separated from USSR (Aydın
& Zortuk, 2014). A hybrid multi-objective meta-heuristic algorithm for obtaining Pareto
optimal solutions of the location routing problem minimising both total cost and total
environmental effect applied in Mohammadi, Razmi, and Tavakkoli-Moghaddam (2013).
Jureniene and Radzevicius (2014) compared three cultural heritage management models
which are generally used in the world UNESCO practices. Iqbal, Choudhry, Ahsan Ali,
and Tamošaitienė (2015) analysed risk management in construction projects in Pakistan.
Two types of risk management techniques were considered: preventive techniques and
remedial techniques. A model of assessing contaminated sites in Lithuania is proposed by
Vasarevičius, Kadūnas, and Baltrėnaitė (2013). Calculations of the level of contamination
have been based on statistical analysis and experience gained by the EU countries.
Lich and Tournemaine (2013) developed an endogenous growth model with human
capital accumulation and pollution unequally spread across geographical locations when
individuals must decide where to live. It was found that individuals prefer a greater level of
consumption and leisure but lower growth and environmental quality than those which are
possible to achieve. Yamashita, Matsuura, and Nakajima (2014) revealed the agglomeration
effects of multinational firms on the location decisions of first-time Japanese manufacturing
investors in China by calculating the conditional and mixed logit estimates. Bryson and
Ronayne (2014) analysed the British textile industry since the 1960s highlighting dein-
dustrialisation and the transfer of production and employment to newly industrialised
regions. Yang, Luo, and Law (2014) reviewed past literature on hotel location models and
provided future research directions related the development of more sophisticated hotel
location models and the use of Geographic Information System in hotel location analy-
sis. Determining the locations of facilities for prepositioning supplies to be used during a
disaster is analysed in Akgün, Gümüşbuğa, and Tansel (2015). A non-linear p-centre type
facility location model to minimise maximum risk is developed in the article. Jaskowski,
Sobotka, and Czarnigowska (2014) proposed mixed binary linear programming models
for the contractor’s logistic decisions assessment.
From the great variety of decision evaluation models applied for solution of location
alternatives problems we would like to distinguish Multiple Criteria Decision Making
(MCDM) or Multiple Attribute Decision Making (MADM) methods. These methods are
naturally suitable for solving the problem of selection an optimal alternative having the set
of criteria which are sometimes conflicting with each other. Furthermore, respective fuzzy
modifications of MCDM methods allow solving such problems under vague conditions.
Various MCDM methods were used for solving the location problems recently. Brauers and
Zavadskas (2008) proposed Multiple Objectives Optimisation by Ratio Analysis (MOORA)
method to resolve location problems. Rikhtegar et al. (2014) applied Analytic Network Process
(ANP) and fuzzy Simple Additive Weight (F-SAW) to formulate the environmental risks pertain-
ing to mining projects. Turskis and Zavadskas (2010) proposed fuzzy additive ratio assessment
(ARAS‐F) method to select the most suitable site for logistic centre among a set of alternatives.
Shariati, Yazdani-Chamzini, Salsani, and Tamosaitiene (2014) considered technical, economic
and environmental factors for waste dump site selection by using the ARAS based group deci-
sion-making fuzzy Group Additive Ratio Assessment (GARAS) technique. Hashemkhani
Zolfani, Aghdaie, Derakhti, Zavadskas, and Varzandeh (2013) presented the new hybrid MCDM
model: Stepwise Weight Assessment Ratio Analysis (SWARA) and Weighted Aggregated Sum
52 A. Krylovas et al.

Product Assessment (WASPAS) methods in shopping mall locating. Bagočius, Zavadskas, and
Turskis (2014a) compared three different MCDM methods: SAW, fuzzy Technique for Order
Preference by Similarity to Ideal Solution (TOPSIS) and Complex Proportional Assessment
(COPRAS) for selecting a location for a liquefied natural gas terminal in the Eastern Baltic Sea.
Zavadskas, Turskis, and Bagočius (2015) applied a combination of Analytic Hierarchy Process
(AHP) and ARAS-F methods for selection of a deep-water port. Yazdani-Chamzini, Yakhchali,
and Zavadskas (2012) used integrated model based on fuzzy AHP and fuzzy TOPSIS for mining
method selection under the uncertainty. Choosing logistics freight centre locations with Fuzzy
Preference Ranking Organisation METHod for Enrichment Evaluation (F-PROMETHEE) is
carried out by Elevli (2014). Sánchez-Lozano, Antunes, García-Cascales, and Dias (2014) pro-
posed the use of Geographic Information System (GIS) and Elimination and Choice Translating
Reality (ELECTRE-TRI) method to identify the best plots suitable for installing photovoltaic
solar farms.
Lee (2014) combined DEMATEL-based Analytic Network Process (DANP) and
VIšekriterijumsko KOmpromisno Rangiranje (VIKOR) methods for selection of location
real estate brokerage services. PROMETHEE is used by Ishizaka and Nemery (2013) for
the site location where partners are working and sharing resources together. AHP and SAW
methods were applied for selection of rural building sites by Jeong, García-Moruno, and
Hernández-Blanco (2013). Dheena and Mohanraj (2011) used Ordered Weighted Averaging
(OWA) operators with maximal entropy for aggregating fuzzy data. Ekmekçioğlu, Kaya,
and Kahraman (2010) proposed AHP and modified fuzzy TOPSIS methodology for selec-
tion the site for municipal solid waste. For the same problem Aragonés-Beltrán, Pastor-
Ferrando, García-García, and Pascual-Agulló (2010) proposed ANP and AHP methods.
Rezaeiniya, Hashemkhani Zolfani, and Zavadskas (2012) developed the hybrid of MCDM
methods – ANP and COPRAS-G – for greenhouse locating. The article by Lee (2015) pro-
poses a strengths, weaknesses, opportunities, and threats – fuzzy analytic network process
(SWOT-FANP) analysis of location selection for a second tier city in China. MCDM method
WASPAS applied for selection and ranking of the feasible location areas of wind farms in
Bagočius, Zavadskas, and Turskis (2014b). The performance of a site assessment model for
the property management company by the integrated TOPSIS method and the signal-to-
noise ratio is proposed by Lin and Pan (2014). Tamošaitienė, Šipalis, Banaitis, and Gaudutis
(2013) analysed models for the assessment of the location of high-rise buildings. SWOT
and SAW methods combination applied for comparison of visions of urban development.
Hsueh, Lee, and Chen (2013) used the Delphi method, fuzzy logic, and AHP (DFAHP) as
a risk assessment model to redevelop derelict public buildings.
The alternative to MCDM approach is Cost-benefit analysis (CBA), where the total
expected cost and the total expected benefit are calculated and compared for each alter-
native (Rietveld, van Binsbergen, Schoemaker, & Peeters, 1998). The debate on CBA and
multicriteria analysis tends to regard these two approaches as complementary rather than
competitive analytical tools.
In addition to the above methods for selection the best alternative of several possible
options the new KEMIRA method is proposed by Krylovas, Zavadskas, Kosareva, and
Dadelo (2014). Frequently there are situations when solving MCDM tasks there is a need
not only to rank alternatives certified by the entirety of criteria. The essence of KEMIRA
method can be formulated as follows: if we can logically distinguish few subsets of criteria,
then for setting criteria weights it is important to include interactions between subsets of
 Economic Research-Ekonomska Istraživanja  53

criteria. KEMIRA method – only one of the possible ways of balancing criteria weights.
In our knowledge it is a new approach and it hasn’t more bibliography. Other balancing
techniques are also possible – one of them is referred to in Dadelo et al. (2014).
At the first stage of the MCDM problem solution it is proposed to determine a priority
of the criteria (indicators) on the basis of expert assessments. For this purpose, Kemeni
median method (Dadelo, Krylovas, Kosareva, Zavadskas, & Dadeliene, 2014) is applied.
Consequently, it becomes clear which criterion will have lower weight and which will have
greater weight. It should be noted that only during this phase criteria priorities determined
by the experts are used. Next, criteria weights must be determined. In most articles criteria
weights are established either by objective, or by subjective methods in both cases using
expert opinions. In this article, we use the Indicator Rank Accordance method presented
for determining weights of the criteria at the second stage of the problem solution. This
method is based entirely on calculations and does not require expert assessments. On the
one hand, decisions must be made based on various expert judgments summary. On the
other hand, the aim is to ensure that decisions made on the basis of the separate groups of
evaluation criteria to be compatible with each other. Therefore, we are solving an additional
optimisation problem – maximisation of criteria interaction. When weights of the criteria
are determined, at the third stage the objective function value is calculated for each alter-
native and the alternatives are ranked according to the objective function values. Usually
this function is a weighted average of all the criteria values calculated with weights, found
by Indicator Rank Accordance method. It should be mentioned that KEMIRA is designed
for determining criteria weights. Method can be applied together with any other method
of establishing priority of the alternatives – SAW, TOPSIS, WASPAS, etc.
Later in this article a concrete example of KEMIRA-M decision method application in
solving the task of construction site choice for non-hazardous waste incineration plant is
shown. The data used are of the Vilnius city.

2. MCDM problem under investigation

Construction site alternatives for non-hazardous waste incineration plant in Vilnius city
were analysed in Turskis, Lazauskas, and Zavadskas (2012). A problem of siting the waste
incineration plant is a complex process which includes social, economic, urban and techno-
logic factors. With the objective to highlight aspects of the application of KEMIRA method,
we use less of data from the article by Turskis et al. (2012). We selected four engineering
factors, which form one separate group of criteria X = (x1 , x2 , x3 , x4 ): x1 - distance to cen-
tralised heating network mains (∅400), km, x2 - distance to the high-pressure (12 bar)
gas supply pipeline (Ø150), km, x3 - distance to 110 kW electric supply networks, km,
x4 - distance to water supply networks (Ø110), km. Engineering factors include a part of
investments required for project development. For initiation of incineration process or its
maintenance, fossil fuel or electric energy will be provided. It is clear that all these factors
will be treated due to their minimisation, i.e., criteria, whose preferable values are minimal.
Urban and social factors in our investigation are combined in the second group of criteria
Y=(y1, y2, y3): y1 - distance to the Vilnius city centre, km, y2 - average number of people
living on the territory of the determined alternative, 1 km2, y3 - usable area of apartments
owned by people living in the territory intended for project implementation, m2. Urban and
social factors are maybe the most controversial. The distance to the city centre evaluates the
location of an incineration plant and considers the possible impact on architecture of the
54 A. Krylovas et al.

Table 1. Alternative assessment factors.

Engineering factors
x1 Distance to centralised heating network mains (∅400), km
x2 Distance to the high-pressure (12 bar) gas supply pipeline (Ø150), km
x3 Distance to 110 kW electric supply networks, km
x4 Distance to water supply networks (Ø110), km
Urban and social factors
y1 Distance to the Vilnius city centre, km
y2 Average number of people living on the territory of the determined alternative, 1 km2
y3 Usable area of apartments owned by people living in the territory intended for project implementation, m2
Source: Turskis et al., 2012.

Table 2. Significance of factors (criteria) determined by five representatives of concerned groups.

Expert group x1 x2 x3 x4 y1 y2 y3
1 1 3 4 2 2 3 1
2 1 3 4 2 1 2 3
3 1 2 3 4 3 1 2
4 1 2 4 3 3 1 2
5 1 2 4 3 1 3 2
Source: Turskis et al., 2012.

city and the resulting problems such as noise, odour or aesthetic view. To avoid all these
negative factors construction site of waste incineration plant should be located as far away
from the city centre as possible. Urban and social factors represent the general evaluation
of conflicting public interests. According to the public opinion survey it was considered
that the site, selected for the Project implementation should have the smallest density of
population. Assessing the Project from the point of view of the state or a private investor,
it is rational to construct a power plant in a densely populated territory to ensure energy
needs. Usable area of apartments owned by citizens shows the area to be supplied with the
energy generated by the new energy facility. In our research all urban and social factors
will be treated due to their maximisation, as it was done in Turskis et al. (2012), i.e. criteria,
whose preferable values are maximal. Alternative assessment factors are proposed in Table 1.
Requirements were formulated by five representatives of concerned groups: citizens,
potential investors, specialists of environmental protection, architects and construction
contractors. Representatives of the above-state groups assessed the stated factors priorities
within each group. Results are given in the Table 2. The lower number in the Table 2 rep-
resents the higher rank (priority) of corresponding criterion.
Note, that priorities of factors xl, ymwere determined independently and separately for
both groups of criteria – engineering and urban/social by non-quantitative assessments
only comparing them to each other. For example, five experts sorted criteria within the
first group as follows:

(E1) x1 ≻ x4 ≻ x2 ≻ x3 ⎤
⎥
(E2) x1 ≻ x4 ≻ x2 ≻ x3 ⎥
(E3) x1 ≻ x2 ≻ x3 ≻ x4 ⎥. (1)
⎥
(E4) x1 ≻ x2 ≻ x4 ≻ x3 ⎥
⎥
(E5) x1 ≻ x2 ≻ x4 ≻ x3 ⎦
 Economic Research-Ekonomska Istraživanja  55

Table 3. Specification of initial values of criteria for seven alternative waste incineration plants sitting in
the Vilnius city.
Criteria
Alternatives x1 x2 x3 x4 y1 y2 y3
a1 1.5 0.6 2.5 1.37 9.26 3188.6 55,269
a2 3.5 1.2 4.5 0.5 8.64 497.5 9,327
a3 0.8 0.5 3 0.1 6.44 2,484 50,798
a4 4.8 1.2 1,6 2 11.19 2,676 56,206
a5 5.5 1 1.6 0.3 5.9 3,291 66,807
a6 0.6 0.7 2 0.6 6.09 6,490 132,136
a7 0.3 0.4 2 0.6 5.72 5946.7 123,314
Source: Turskis et al., 2012.

Table 4. Specification of transformed criteria values treated due to their maximisation.

Criteria
Alternatives x1 x2 x3 x4 y1 y2 y3
a1 0.667 1.667 0.4 0.730 9.26 3188.6 55,269
a2 0.286 0.833 0.222 2.0 8.64 497.5 9,327
a3 1.250 2.0 0.333 10.0 6.44 2,484 50,798
a4 0.208 0.833 0.625 0.5 11.19 2,676 56,206
a5 0.182 1.0 0.625 3.333 5.9 3,291 66,807
a6 1.667 1.429 0.50 1.667 6.09 6,490 132,136
a7 3.333 2.50 0.50 1.667 5.72 5946.7 123,314
Source: calculated by the authors.

Table 5. Normalised criteria values for seven alternative waste incineration plant sitting in the Vilnius city.
Criteria
Alternatives x1* x2* x3* x4* y1* y2* y3*
a1 0.154 0.5 0.441 0.024 0.647 0.449 0.374
a2 0.033 0,0 0.0 0.158 0.534 0.0 0.0
a3 0.339 0.7 0.276 1.0 0.132 0.331 0.338
a4 0.008 0.0 1.0 0.0 1.0 0.364 0.382
a5 0.0 0.1 1.0 0.298 0.033 0.466 0.468
a6 0.471 0.357 0.690 0.123 0.068 1.0 1.0
a7 1.0 1.0 0.690 0.123 0.0 0.909 0.928
Source: created by the authors.

In the article by Turskis et al. (2012), the five expert’s quantitative information is processed
using fuzzy sets. In this article, we will use the aggregated data for determination of the order
found by each expert. Dividing criteria into groups makes it easier to rank them, because
experts must rank shorter lists of criteria (factors) – four and three factors respectively.
Seven local assessments of alternative sites a1 - a7 located in densely populated urban and
industrial development areas have been evaluated. In the Table 3 numerical estimates of
‖ ‖ ‖ ‖
the matrix factors for all the alternatives are presented: X(a) = ‖xij ‖ , Y (a) = ‖yij ‖ .
‖ ‖7x4 ‖ ‖7x3
At the next step all criteria values must be/transformed to the factors treated due to their
maximisation, i.e., for x1-x4 inverse values (1 xi) must be calculated. Since y1-y3 are treated
due to their maximisation, we do not accomplish any transformations of these factors.
Transformed values are presented in Table 4.
Further, all criteria values are normalised according to the formulas:
56 A. Krylovas et al.

xj(i) − xmin
(i)
yj(i) − ymin
(i)

xj(i)∗ = (i) (i)

, yj(i)∗ = (i) (i)
, i = 1, 2, ..., 7. (2)
xmax − xmin ymax − ymin

Table 5 presents normalised criteria values for 7 alternatives.

Normalised criteria values are belonging to the interval [0; 1]. Furthermore, all criteria
values in Table 5 are due to their maximisation.

3. Determining priority of criteria by Kemeny median method

The main idea of KEMIRA method is to construct the median, i.e., generalised expert opin-
ion of (1) estimates. It would be one or few of all 4! = 24 possible rankings, since there are 24
permutations of ranks 1, 2, 3, 4. Then on the basis of the median we form weighted average:

XWX = 𝜛x1 x1 + 𝜛x2 x2 + 𝜛x3 x3 + 𝜛x4 x4 , WX = (𝜛x1 , 𝜛x2 , 𝜛x3 , 𝜛x4 ). (3)

For example, if the median priority is an order x1 ≻ x4 ≻ x2 ≻ x3, then the weights in (3)
must satisfy the following constraints:
𝜛x1 ≥ 𝜛x4 ≥ 𝜛x2 ≥ 𝜛x3 , 𝜛x1 +𝜛x4 +𝜛x2 + 𝜛x3 = 1. (4)
The weighted average of the other factor YWY is constructed by analogy. Weights ϖx, ϖyare
chosen so that the criteria XWX and YWY values are as close as possible. KEMIRA method
algorithm consists of the following steps:
(1) All criteria represent a benefit, i.e., the bigger is a value, the better is the respective
alternative. If criterion
/ xi is representing lost, when the lower value is better, then
inverse value (1 xi) must be calculated.
(2) Normalisation of elements of decision-making matrices X(a), Y(a) by formula (2).
All criteria values after normalisation belong to the interval [0; 1].
(3) Determining median priorities of criteria X and Y (obtaining one or few medians).
(4) The weights WX, WY are chosen so that they satisfy the constraints (4) and the
difference of weighted averages XWX and YWY for real data is minimised at all the
alternatives. In order to achieve this goal optimisation problem described below
is solved.
(5) The best alternative is the one for which the sum of weighted averages XWX + YWY
reach its maximum value.

3.1. Factor X median construction

According to KEMIRA method at the first stage Kemeny median method is applied for the
data in the Table 2 for establishing priority of criteria separately and independently in both
groups - engineering and urban/social factors. Priority of criteria X = (x1, x2, x3, x4) and Y =
(y1, y2, y3) was estimated by five experts (see data in Table 2). Notice, that all experts identified
the first factor as the most important one. Meanwhile, expert opinions on importance of the
other factors differ. We must ascertain generalised expert opinion. Each of the five experts’
formations can be seen as a simple oriented graph (Krylovas, 2009). For example, first expert
 Economic Research-Ekonomska Istraživanja  57

formation x1 ≻ x4 ≻ x2 ≻ x3 is expressed by graph R(1) = {(1;2), (1;3);(1;4);(4;2), (4;3), (2;3)}

which is depicted in Fig. 1. ( )
The same graph can be written in the form of { the square matrix A(r) = aij ,
4x4
0, if xi ≺ xj
r = 1, 2, ..., 5 which elements are as follows: aij = . Diagonal elements
1, if xi ≻ xj
of this matrix are equal to zero: xii = 0, i = 1, 2, 3, 4. Moreover, aij = 1 - aji, i ≠ j. For
example, significance of factors x1 - x4 set by first and second experts (E1) - (E2) can be
⎛ 0 1 1 1 ⎞
⎜ 0 0 1 0 ⎟
described by matrices: A(1) = A(2) =⎜ ⎟, other experts matrices are as follows:
⎜ 0 0 0 0 ⎟
⎜ ⎟
⎝ 0 1 1 0 ⎠
⎛ 0 1 1 1 ⎞ ⎛ 0 1 1 1 ⎞
⎜ 0 0 1 1 ⎟ ⎜ ⎟
A(3) =⎜ ⎟, A(4) = A(5) = ⎜ 0 0 1 1 ⎟.
⎜ 0 0 0 1 ⎟ ⎜ 0 0 0 0 ⎟
⎜ ⎟ ⎜ ⎟
⎝ 0 0 0 0 ⎠ ⎝ 0 0 1 0 ⎠
The measure of distance between two different ranking orders described by matrices A(r)
and A(s) is defined by formula (5):
∑
n
∑
n
| (r) (s) |
(r) (s)
𝜌(R , R ) = 𝜌(A , A ) =(r)
|aij − aij |.
(s)
(5)
| |
i=1 j=1

Here n is number of factors x1, x2, ..., xn (n = 4). Formula (5) calculates sum of absolute
values of corresponding elements differences. For example,
𝜌(A(1) , A(3) ) = |0 − 0| + |1 − 1| + |1 − 1| + |1 − 1|+
|0 − 0| + |0 − 0| + |1 − 1| + |0 − 1|+
|0 − 0| + |0 − 0| + |0 − 0| + |0 − 1|+
|0 − 0| + |1 − 0| + |1 − 0| + |0 − 0| = 4 .

Suppose that S experts established priorities which can be described by matrices

A(1), A(2), ..., A(S) (S = 5). Most consistent with these estimates will be priority described by
∑S
matrixA(M), which is called median. For the median A = A(M) the sum 𝜌(A, A(j) ) is gain-
j=1
‖ ‖
ing its minimum value. Theoretically, we must search the median matrix A(M) = ‖aij ‖
‖ ‖4x4
within all 24 possible matrices. But we can narrow down the search area. Notice that
if elements of all five matrices with the corresponding indices i and j are equal, then
⎛ 0 1 1 1 ⎞
⎜ 0 0 1 a ⎟
aij(M) = aij(1) = aij(2) = ... = aij(5). Then A(M) = ⎜ 24 ⎟
. We have only 2 unknown
⎜ 0 0 0 a34 ⎟
⎜ ⎟
⎝ 0 a42 a43 0 ⎠
elements of median matrix, because the identities aij + aji = 1, i ≠ j are valid for the simple
58 A. Krylovas et al.

Figure 1. Simple oriented graph describing priority of criteria X by first expert x1 ≻ x4 ≻ x2 ≻ x3. Source:
created by the authors.

oriented graph matrix (Krylovas, 2009). Therefore a42 = 1 - a24, a43 = 1 - a34. Thus, we can
search the median within these 4 matrices:

⎛ 0 1 1 1 ⎞ ⎛ 0 1 1 1 ⎞ ⎛ 0 1 1 1 ⎞ ⎛ 0 1 1 1 ⎞
⎜ 0 0 1 1 ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
A� = ⎜ ⎟, A�� = ⎜ 0 0 1 1 ⎟, A��� = ⎜ 0 0 1 0 ⎟, A���� = ⎜ 0 0 1 0 ⎟
⎜ 0 0 0 1 ⎟ ⎜ 0 0 0 0 ⎟ ⎜ 0 0 0 1 ⎟ ⎜ 0 0 0 0 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝ 0 0 0 0 ⎠ ⎝ 0 0 1 0 ⎠ ⎝ 0 1 0 0 ⎠ ⎝ 0 1 1 0 ⎠

Calculate sums (5) for each of these matrices:

∑
5
∑
5
𝜌(A� , A(j) ) = 4 + 4 + 0 + 2 + 2 = 12, 𝜌(A�� , A(j) ) = 2 + 2 + 2 + 0 + 0 = 6,
j=1 j=1
∑
5
∑
5
𝜌(A��� , A(j) ) = 2 + 2 + 2 + 4 + 4 = 14, 𝜌(A���� , A(j) ) = 0 + 0 + 4 + 2 + 2 = 8.
j=1 j=1

Minimum value of sum (5) is reached for matrix A′′.1 Hence, median value which in the
best way represents all five experts’ opinions is (1, 2, 4, 3) which corresponds to preference
of engineering factors as follows:
x1 ≻ x2 ≻ x4 ≻ x3 . (6)
The highest priority has the distance to centralised heating network mains, then goes dis-
tance to the high-pressure gas supply pipeline, after this goes the distance to water supply
networks, and finally, the lowest priority has distance to electric supply networks. Write
∑4
down the weighted average of type (3) for each alternative XWX (a) = 𝜛xj xj (a). Here the
j=1
weights 𝜛x1 , 𝜛x2 , 𝜛x3 , 𝜛x4 must satisfy the following constraints:
𝜛x1 ≥ 𝜛x2 ≥ 𝜛x4 ≥ 𝜛x3 ≥ 0, 𝜛x1 +𝜛x2 +𝜛x3 + 𝜛x4 = 1. (7)

3.2. Factor Y median construction

The same procedure is carried out with other group of factors – urban and social factors
y1, y2, y3. Five matrices, describing experts’ priorities (2,3,1), (1,2,3), (3,1,2), (3,1,2) and
(1,3,2) (see Table 2) are as follows:
 Economic Research-Ekonomska Istraživanja  59

Table 6. Permutations of three ranks describing three factors prioritisation options, corresponding ma-
trices and values of function F.
Permutation (1, 2, 3) (1, 3, 2) (2, 1, 3) (2, 3, 1) (3, 1, 2) (3, 2, 1)
Priority of y1 ≻ y2 ≻ y3 y1 ≻ y3 ≻ y2 y2 ≻ y1 ≻ y3 y3 ≻ y1 ≻ y2 y2 ≻ y3 ≻ y1 y3 ≻ y2 ≻ y1
factors
Matrix A ⎛ 0 1 1 ⎞ ⎛ 0 1 1 ⎞ ⎛ 0 0 1 ⎞ ⎛ 0 1 0 ⎞ ⎛ 0 0 0 ⎞ ⎛ 0 0 0 ⎞
⎜ 0 0 1 ⎟ ⎜ 0 0 0 ⎟ ⎜ 1 0 1 ⎟ ⎜ 0 0 0 ⎟ ⎜ 1 0 1 ⎟ ⎜ 1 0 0 ⎟
⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎝ 0 0 0 ⎠ ⎝ 0 1 0 ⎠ ⎝ 0 0 0 ⎠ ⎝ 1 1 0 ⎠ ⎝ 1 0 0 ⎠ ⎝ 1 1 0 ⎠
F 14 16 16 14 14 16
Source: created by the authors.

⎛ 0 1 0 ⎞ ⎛ 0 1 1 ⎞ ⎛ 0 0 0 ⎞ ⎛ 0 1 1 ⎞
(1) ⎜ ⎟ (2) ⎜ ⎟ (3) (4) ⎜ ⎟ (5) ⎜ ⎟
A = ⎜ 0 0 0 ⎟, A = ⎜ 0 0 1 ⎟, A = A = ⎜ 1 0 1 ⎟, A = ⎜ 0 0 0 ⎟
⎜ 1 1 0 ⎟ ⎜ 0 0 0 ⎟ ⎜ 1 0 0 ⎟ ⎜ 0 1 0 ⎟
⎝ ⎠ ⎝ ⎠ ⎝ ⎠ ⎝ ⎠

Next we search median matrix A(M) within 6 possible matrices, representing differ-
ent priorities of three factors. In the last row of Table 6 calculated values of func-
∑5
∑3 3
∑ � (r) (s) �
tion F = 𝜌(A, A(i) ) are given. Here 𝜌(A(r) , A(s) ) = �aij − aij �. For example, if
i=1 i=1 j=1 � �
⎛ 0 1 1 ⎞
⎜ ⎟ ∑ 5
A = ⎜ 0 0 1 ⎟, 𝜌(A, A(j) ) = 4 + 0 + 4 + 4 + 2 = 14.
⎜ 0 0 0 ⎟ j=1
⎝ ⎠
Median components priority is not unique is this case. There are three median values
for which functions F value equals to 14: (1, 2, 3), (2, 3, 1), (3, 1, 2). Respective preferences
of urban/social factors are as follows:
y1 ≻ y2 ≻ y3 , y3 ≻ y1 ≻ y2 , y2 ≻ y3 ≻ y1 . (8)
For the purpose of demonstration of our method let’s choose second median priority of
urban/social factors: y3 ≻ y1 ≻ y2. Thus, the highest priority has usable area of apartments
owned by people living in the territory intended for project implementation, after goes
the distance to Vilnius city centre and the lowest priority has the number of people liv-
ing in the territory of the determined alternative. Then we calculate the weighted average
∑ 3
YWY (a) = 𝜛yj yj (a) with the weights satisfying conditions:
j=1
𝜛y3 ≥ 𝜛y1 ≥ 𝜛y2 ≥ 0, 𝜛y1 + 𝜛y2 + 𝜛y3 = 1. (9)

4. Calculation of weights by modified Indicator Rank Accordance method

and ranking the alternatives
Now when priority of criteria is clear, the next step is finding weighted coefficients
ϖxi, ϖyi ∊ [0; 1] which satisfy conditions (7) and (9). Further, the decision on ranking
the alternatives will be made according to the values of sums of linear combinations
XWX (a) + YWY (a), calculated for each alternative a1, a2, ..., a7:
60 A. Krylovas et al.

Table 7. Weights 𝜛xi combinations satisfying condition 𝜛x1 ≥ 𝜛x2 ≥ 𝜛x4 ≥ 𝜛x3 ≥ 0.
No. 𝜛x3 𝜛x4 𝜛x2 𝜛x1 Nr. 𝜛x3 𝜛x4 𝜛x2 𝜛x1
1 0 0 0 1 13 0 0.1 0.4 0.5
2 0 0 0.1 0.9 14 0 0.2 0.3 0.5
3 0 0 0.2 0.8 15 0.1 0.1 0.3 0.5
4 0 0.1 0.1 0.8 16 0.1 0.2 0.2 0.5
5 0 0 0.3 0.7 17 0 0.2 0.4 0.4
6 0 0.1 0.2 0.7 18 0.1 0.1 0.4 0.4
7 0.1 0.1 0.1 0.7 19 0 0.3 0.3 0.4
8 0 0 0.4 0.6 20 0.1 0.2 0.3 0.4
9 0 0.1 0.3 0.6 21 0.2 0.2 0.2 0.4
10 0 0.2 0.2 0.6 22 0.1 0.3 0.3 0.3
11 0.1 0.1 0.2 0.6 23 0.2 0.2 0.3 0.3
12 0 0 0.5 0.5
Source: created by the authors.

Table 8. Weights 𝜛yi combinations satisfying condition 𝜛y3 ≥ 𝜛y1 ≥ 𝜛y2 ≥ 0.

No. 𝜛y2 𝜛y1 𝜛y3 Nr. 𝜛y2 𝜛y1 𝜛y3
1 0 0 1 8 0.1 0.3 0.6
2 0 0.1 0.9 9 0.2 0.2 0.6
3 0 0.2 0.8 10 0 0.5 0.5
4 0.1 0.1 0.8 11 0.1 0.4 0.5
5 0 0.3 0.7 12 0.2 0.3 0.5
6 0.1 0.2 0.7 13 0.2 0.4 0.4
7 0 0.4 0.6 14 0.3 0.3 0.4
Source: created by the authors.

∑
4
∑
3
XWX (a) = 𝜛xi xi∗ (a), YWY (a) = 𝜛yi yi∗ (a) (10)
i=1 i=1

xi∗ (a) and yi∗ (a) are normalised elements of decision-making matrix, whose all criteria are
due to their maximum value. The idea of modified Indicator Rank Accordance method
is to select such values of coefficients WX = (𝜛x1 , 𝜛x2 , 𝜛x3 , 𝜛x4 ) and WY = (𝜛y1 , 𝜛y2 , 𝜛y3 )
which guarantee the proximity of values XWX (a) and YWY (a):
∑| |
F(X, Y ) = min |XWX (a) − YWY (a)| (11)
WX ,WY | |
a

Note, that in the original KEMIRA method sum of squares of ranks differences are mini-
mised instead of minimising the differences of functions XWX (a) and YWY (a) original values
(11). The other difference of KEMIRA-M from KEMIRA method is that sum (11) is over
all the alternatives, while in the original KEMIRA sum is over only the best alternatives.
However, if there are only few criteria in each criteria group (4 and 3 in our problem),
the number of coefficients WX and WY collections minimising objective function usually
became very big and it is impossible to find the unique solution of the problem. The variety
of function (11) values is much wider than when using sum of squares of ranks differences.
We will show that the solution of optimisation problem (11) in this case is unique. We
will search for the approximate solution of the problem constructing finite sets of options
under consideration. Weights WX are constructed in the following way. Suppose that WXi
 Economic Research-Ekonomska Istraživanja  61

Table 9. Values of function F(X, Y) for all possible weights combinations.

1 2 3 4 5 6 7
1 1.696 1.750 1.869 1.756 1.988 1.877 2.107
2 1.696 1.757 1.876 1.765 1.995 1.884 2.114
3 1.697 1.764 1.883 1.772 2.002 1.891 2.121
4 1.738 1.742 1.861 1.750 1.980 1.869 2.099
5 1.698 1.771 1.890 1.779 2.009 1.898 2.128
6 1.739 1.749 1.868 1.757 1.987 1.876 2.106
7 1.510 1.458 1.577 1.466 1.696 1.585 1.815
8 1.698 1.778 1.897 1.786 2.016 1.905 2.135
9 1.739 1.756 1.875 1.746 1.994 1.883 2.113
10 1.924 1.755 1.852 1.759 1.972 1.860 2.091
11 1.511 1.465 1.584 1.473 1.703 1.592 1.822
12 1.698 1.785 1.904 1.793 2.023 1.912 2.142
8 9 10 11 12 13 14
1 1.996 1.885 2.227 2.115 2.004 2.123 2.012
2 2.003 1.892 2.234 2.122 2.011 2.130 2.019
3 2.010 1.899 2.241 2.130 2.018 2.137 2.026
4 1.988 1.877 2.218 2.107 1.996 2.115 2.004
5 2.017 1.906 2.248 2.136 2.025 2.114 2.033
6 1.995 1.884 2.225 2.114 2.003 2.122 2.011
7 1.704 1.593 1.935 1.823 1.712 1.831 1.720
8 2.024 1.913 2.255 2.143 2.032 2.151 2.040
9 2.002 1.891 2.232 2.121 2.010 2.129 2.018
10 1.980 1.868 2.210 2.100 1.988 2.107 1.996
11 1.711 1.600 1.942 1.830 1.719 1.838 1.727
12 2.031 1.920 2.262 2.151 2.039 2.158 2.047
1 2 3 4 5 6 7
13 1.740 1.763 1.882 1.771 2.000 1.890 2.120
14 1.924 1.762 1.859 1.766 1.979 1.867 2.098
15 1.511 1.472 1.591 1.480 1.710 1.599 1.829
16 1.696 1.533 1.569 1.537 1.688 1.577 1.807
17 1.925 1.769 1.866 1.773 1.986 1.874 2.105
18 1.511 1.479 1.598 1.487 1.717 1.606 1.836
19 2.110 1.936 1.856 1.941 1.963 1.860 2.083
20 1.696 1.540 1.576 1.544 1.695 1.584 1.814
21 1.467 1.311 1.285 1.316 1.404 1.293 1.523
22 1.881 1.708 1.634 1.712 1.680 1.638 1.799
23 1.468 1.318 1.292 1.322 1.411 1.300 1.530
8 9 10 11 12 13 14
13 2.009 1.898 2.239 2.128 2.017 2.136 2.025
14 1.987 1.875 2.217 2.106 1.995 2.114 2.003
15 1.718 1.607 1.949 1.837 1.726 1.845 1.734
16 1.696 1.585 1.926 1.815 1.704 1.823 1.712
17 1.994 1.883 2.224 2.113 2.002 2.121 2.010
18 1.725 1.614 1.956 1.844 1.733 1.852 1.741
19 1.971 1.864 2.202 2.091 1.979 2.099 1.987
20 1.703 1.592 1.933 1.822 1.711 1.830 1.917
21 1.412 1.301 1.701 1.531 1.420 1.539 1.428
22 1.688 1.642 1.918 1.807 1.696 1.815 1.704
23 1.419 1.308 1.728 1.538 1.427 1.546 1.435
Source: calculated by the authors.

are non-negative integers 0, 1, 2, ..., 10 and 𝜛xi = WXi ∕10. Then all 23 weights combinations
satisfying conditions (7) are written in the Table 7.
Then by analogy we construct 14 weights 𝜛yi = WYi ∕10 combinations satisfying condi-
tions (9). These weights are presented in the Table 8.
Calculate and write in the Table 9 values of function F(X, Y)for all 23 ⋅ 14 = 322cases.
Function F(X, Y) gains its minimum value 1.285. First solution with factor X median
(6) and factor Y median y3 ≻ y1 ≻ y2was reached with the following coefficients values:
62 A. Krylovas et al.

Table 10. Final ranking of alternatives for setting the optimal solution.
ϖx, ϖy 0.2 0.2 0.2 0.4 0 0.2 0.8
Alternative x3 x4 x2 x1 y2 y1 y3 XW YW XW + Y W Rank
X Y X Y

a1 0.4414 0.0242 0.5000 0.1538 0.4491 0.6472 0.3741 0.2547 0.4287 0.6834 5
a2 0.0000 0.1579 0.0000 0.0330 0.0000 0.5338 0.0000 0.0448 0.1068 0.1515 7
a3 0.2759 1.000 0.7000 0.3389 0.3315 0.1316 0.3377 0.5307 0.2965 0.8272 3
a4 1.0000 0.0000 0.0000 0.0084 0.3635 1.0000 0.3817 0.2034 0.5054 0.7087 4
a5 1.0000 0.2982 0.1000 0.000 0.4662 0.0329 0.4680 0.2796 0.3810 0.6607 6
a6 0.6897 0.1228 0.3571 0.4712 1.0000 0.0676 1.000 0.4224 0.8135 1.2359 2
a7 0.6897 0.1228 1.000 1.000 0.9093 0.0000 0.9282 0.7625 0.7425 1.5050 1
Source: calculated by the authors.

𝜛x1 = 0.4, 𝜛x2 = 𝜛x4 = 𝜛x3 = 0.2, 𝜛y3 = 0.8, 𝜛y1 = 0.2, 𝜛y2 = 0. (12)
By analogy calculations were repeated for other two medians of urban/social factors group.
For the median y1 ≻ y2 ≻ y3 minimum value of function F(X, Y) = 1.547 was obtained with
the following coefficients values:
𝜛x1 = 0.4, 𝜛x2 = 𝜛x4 = 𝜛x3 = 0.2, 𝜛y1 = 0.4, 𝜛y2 = 𝜛y3 = 0.3 and finally, for
the median y2 ≻ y3 ≻ y1minimum value was F(X, Y) = 1.317 with the coefficients
𝜛x1 = 0.4, 𝜛x2 = 𝜛x4 = 𝜛x3 = 0.2, 𝜛y2 = 𝜛y3 = 0.4, 𝜛y1 = 0.2. The lowest value is 1.285
and calculated weights are (12). Further we must rank our alternatives by calculating values
of both functions (10) and their sum XWX (a) + YWY (a) and rank alternatives according to
its value. Final results presented in the Table 10.
Final ranks assigned to the alternatives according to the values XWX (a) + YWY (a) are in
the last column of Table 10. Thus ranking order of the alternatives is a7 ≻ a6 ≻ a3 ≻ a4 ≻
a1 ≻ a5 ≻ a2. Note that the solution of the same problem obtained in the article by Turskis
et al. (2012) gives slightly different order of alternatives, namely alternatives a4 and a5
switched places.

5. Conclusion and discussion

Modified KEMIRA method is convenient and recommended to use for establishing cri-
teria weights when there are few groups of factors (criteria) and the number of criteria in
each group is small. In the case study we had four engineering factors and three social/
urban factors. For this example, procedure of median search is rather simple especially if
we can restrict our search to the reduced number of possible cases as we did with factor X.
Significance of criteria determined by experts was used. Procedure or setting criteria order
separately in each group is easier, because experts must rank shorter lists of criteria. The
next advantage of the method – using much less information compared with other MCDM
methods. It is enough to have only information about criteria rankings accomplished by few
experts. The demonstrated procedure of problem solving by modified KEMIRA method is
not difficult comparing with other MCDM methods and it is easier to carry out calculations
compared to original KEMIRA method.
The main idea of the method is to create uniform mesh and calculate values of objective
function for all points of this mesh. As the result we obtained the approximate solution.
If it is need to search more accurate solution the mesh should be smaller. Compared with
Turskis et al. (2012) the obtained solution is very similar – only two alternatives switched
 Economic Research-Ekonomska Istraživanja  63

places. The advantage of KEMIRA method is that having much less initial information –
we have operated with seven criteria instead of 10, as in Turskis et al. (2012) – we obtained
analogous result. Therefore, KEMIRA together with weighted sum of criteria values used
for ranking the alternatives is superior in comparison with AHP with ARAS-F method. This
example showed that KEMIRA is suitable for solving MCDM problems in urban planning.
Suitability of the method is based on the fact that it is rather difficult to gather information
for the decision-making matrix and expert evaluation of the criteria. For the same reason
this method is appropriate for application in the fields where information is difficult to
gather or it is rather expensive.
All methods of balancing criteria weights are applicable only under the restrictions
described in the article and for this reason it is difficult to compare them with other meth-
ods precisely.
KEMIRA method can be easily summarised for three types of criteria (we are currently
conducting such studies). For the case when there is only one type of criteria it should be
either some other additional external criteria, or the set of criteria has to be divided to few
groups artificially. Brute force (i.e. the total re-selection) algorithm implementation in Excel
was chosen for this particular task, as these resources are enough and our objective was to
introduce the method so that it could be applied by engineers. In the previous investigation
(Dadelo et el. 2014) we employed the partial re-selection and random search combination.
The authors of the article are currently conducting global and local random search algorithm
implementation with C ++.

Note
1. 
Notice, that matrix A′′′ can’t represent any priority of criteria, because it doesn’t contain the
row with all zero values.

Disclosure statement
No potential conflict of interest was reported by the authors.

References
Akgün, I., Gümüşbuğa, F., & Tansel, B. (2015). Risk based facility location by using fault tree analysis
in disaster management. Omega, 52, 168–179.
Ansar, A. (2013). Location decisions of large firms: Analyzing the procurement of infrastructure
services. Journal of Economic Geography, 13, 823–844.
Aragonés-Beltrán, P., Pastor-Ferrando, J. P., García-García, F., & Pascual-Agulló, A. (2010). An analytic
network Process approach for siting a municipal solid waste plant in the Metropolitan area of
valencia (Spain). Journal of Environmental Management, 91, 1071–1086.
van Asperen, E., & Dekker, R. (2013). Centrality, flexibility and floating stocks: A quantitative
evaluation of port-of-entry choices. Maritime Economics & Logistics, 15, 72–100.
Aydın, N., & Zortuk, M. (2014). Measuring efficiency of foreign direct investment in selected transition
economies with fuzzy data envelopment analysis. Economic Computation And Economic Cybernetics
Studies And Research, 48, 273–286.
Bagočius, V., Zavadskas, E. K., & Turskis, Z. (2014a). Selecting a location for a liquefied natural gas
terminal in the Estern Baltic sea. Transport, 29, 69–74.
64 A. Krylovas et al.

Bagočius, V., Zavadskas, E. K., & Turskis, Z. (2014b). Multi-person selection of the best wind turbine
based on the multi-criteria integrated additive-multiplicative utility function. Journal of Civil
Engineering and Management, 20, 590–599.
Brauers, W. K. M., & Zavadskas, E. K. (2008). Multi-objective optimization in location theory with a
simulation for a department store. Transformations in business & economics, 7, 163–183.
Broll, U., Roldán-Ponce, A., & Wahl, J. E. (2013). Regional investment under uncertain costs of
location. The Annals of Regional Science, 51, 645–657.
Bryson, J. R., & Ronayne, M. (2014). Manufacturing carpets and technical textiles: routines, resources,
capabilities, adaptation, innovation and the evolution of the British textile industry. Cambridge
Journal of Regions, Economy and Society, 7, 471–488.
Dadelo, S., Krylovas, A., Kosareva, N., Zavadskas, E. K., & Dadeliene, R. (2014). Algorithm of
maximizing the set of common solutions for several MCDM problems and it’s application for
security personnel scheduling. International Journal of Computers, Communications & Control,
9, 140–148.
Dheena, P., & Mohanraj, G. (2011). Multicriteria decision-making combining fuzzy set theory, ideal
and anti-ideal points for location site selection. Expert Systems with Applications, 38, 13260–13265.
Ekmekçioğlu, M., Kaya, T., & Kahraman, C. (2010). Fuzzy multicriteria disposal method and site
selection for municipal solid waste. Waste Management, 30, 1729–1736.
Elevli, B. (2014). Logistics freight center locations decision by using Fuzzy-PROMETHEE. Transport,
29, 412–418.
Gerritse, M., & Moreno-Monroy, A. M. (2012). Informal sector and manufacturing location. In search
of the missing links. Spatial Economic Analysis, 7, 179–201.
Gilvear, D. J., Spray, C. J., & Casas-Mulet, R. (2013). River rehabilitation for the delivery of multiple
ecosystem services at the river network scale. Journal of Environmental Management, 126, 30–43.
Hashemkhani Zolfani, S., Aghdaie, M. H., Derakhti, A., Zavadskas, E. K., & Varzandeh, M. H. M.
(2013). Decision making on business issues with foresight perspective; an application of new
hybrid MCDM model in shopping mall locating. Expert Systems with Applications, 40, 7111–7121.
Hsueh, S.-L., Lee, J.-R., & Chen, Y.-L. (2013). DFAHP multicriteria risk assessment model for
redeveloping derelict public buildings. International Journal of Strategic Property Management,
17, 333–346.
Iqbal, S., Choudhry, R. M., Ahsan Ali, K. H., & Tamošaitienė, J. (2015). Risk management in
construction projects. Technological and Economic Development of Economy, 21, 65–78.
Ishizaka, A., & Nemery, P. (2013). A multi-criteria group decision framework for partner grouping
when sharing facilities. Group Decision and Negotiation, 22, 773–799.
Jaskowski, P., Sobotka, A., & Czarnigowska, A. (2014). Decision model for selecting supply sources
of road construction aggregates. Inzinerine Ekonomika-Engineering Economics, (2014), 25, 13–20.
Jeong, J. S., García-Moruno, L., & Hernández-Blanco, J. (2013). A site planning approach for rural
buildings into a landscape using a spatial multi-criteria decision analysis methodology. Land Use
Policy, 32, 108–118.
Jureniene, V., & Radzevicius, M. (2014). Models of cultural heritage management. Transformations
in Business & Economics, 13, 236–256.
Krylovas, A. (2009). Diskrečioji matematika[Discrete mathematics]. Vilnius: Technika.
Krylovas, A., Zavadskas, E. K., Kosareva, N., & Dadelo, S. (2014). New KEMIRA method for
determining criteria priority and weights in solving MCDM problem. International Journal of
Information Technology & Decision Making, 13, 1119–1133.
Lee, W.-S. (2014). A new hybrid MCDM model combining DANP with VIKOR for the selection
of location – real estate brokerage services. International Journal of Information Technology &
Decision Making, 13, 197–224.
Lee, Y.-H. (2015). Navigating SWOT-FANP with GSM method to prioritize the strategic location.
Technological and Economic Development of Economy, 21, 140–163.
Lich, H. K., & Tournemaine, F. (2013). Choice of location, growth and welfare with unequal pollution
exposures. Technological and Economic Development of Economy, 19, S58–S82.
Lin, Y. Y., & Pan, N. H. (2014). Multi period performance assessment model for the site property
management. International Journal of Strategic Property Management, 18, 332–343.
 Economic Research-Ekonomska Istraživanja  65

Mohammadi, M., Razmi, J., & Tavakkoli-Moghaddam, R. (2013). Multi-objective invasive weed
optimization for stochastic green hub location routing problem with simultaneous pick-ups and
deliveries. Economic Computation and Economic Cybernetics Studies and Research, 47, 247–266.
Rezaeiniya, N., Hashemkhani Zolfani, S., & Zavadskas, E. K. (2012). Greenhouse locating based on
ANP-COPRAS-G methods – an empirical study based on Iran. International Journal of Strategic
Property Management, 16, 188–200.
Rietveld, P., van Binsbergen, A., Schoemaker, T., & Peeters, P. (1998). Optimal speed limits for various
types of roads: A social cost benefit analysis for the Netherlands. In R. Roson & K. A. Small (Eds.),
Environment and Transport in Economic Modelling (pp. 206–225). Dordrecht: Kluwer.
Rikhtegar, N., Mansouri, N., Oroumieh, A. A., Yazdani-Chamzini, A., Zavadskas, E. K., & Kildienė,
S. (2014). Environmental impact assessment based on group decision-making methods in mining.
Economic Research-Ekonomska Istraživanja, 27, 378–392.
Sánchez-Lozano, J., Antunes, C., García-Cascales, M., & Dias, L. (2014). GIS-based photovoltaic
solar farms site selection using ELECTRE-TRI: Evaluating the case for Torre Pacheco, Murcia,
Southeast of Spain. Renewable Energy, 66, 478–494.
Shariati, S., Yazdani-Chamzini, A., Salsani, A., & Tamosaitiene, J. (2014). Proposing a new model
for waste dump site selection: Case study of ayerma phosphate mine. Inzinerine Ekonomika-
Engineering Economics, 25, 410–419.
Tamošaitienė, J., Šipalis, J., Banaitis, A., & Gaudutis, E. (2013). Complex model for the assessment
of the location of high-rise buildings in the city urban structure. International Journal of Strategic
Property Management, 17, 93–109.
Turskis, Z., Lazauskas, M., & Zavadskas, E. K. (2012). Fuzzy multiple criteria assessment of construction
site alternatives for non-hazardous waste incineration plant in vilnius city, applying ARAS-F and
AHP methods. Journal of Environmental Engineering and Landscape Management, 20, 110–120.
Turskis, Z., & Zavadskas, E. K. (2010). A new fuzzy additive ratio assessment method (ARAS-F).
Case study: The analysis of fuzzy multiple criteria in order to select the logistic centers location.
Transport, 25, 423–432.
Vasarevičius, S., Kadūnas, K., & Baltrėnaitė, E. (2013). Comparison of requirements for environmental
protection and a model for evaluating contaminated sites in Lithuania. Technological and Economic
Development of Economy, 19, 63–77.
Yamashita, N., Matsuura, T., & Nakajima, K. (2014). Agglomeration effects of inter-firm backward
and forward linkages: Evidence from Japanese manufacturing investment in China. Journal of the
Japanese and International Economies, 34, 24–41.
Yang, Y., Luo, H., & Law, R. (2014). Theoretical, empirical, and operational models in hotel location
research. International Journal of Hospitality Management, 36, 209–220.
Yazdani-Chamzini, A., Yakhchali, S. H., & Zavadskas, E. K. (2012). Using a integrated MCDM
model for mining method selection in presence of uncertainty. Economic Research-Ekonomska
Istraživanja, 25, 869–904.
Zavadskas, E. K., Turskis, Z., & Bagočius, V. (2015). Multi-criteria selection of a deep-water port in
the Eastern Baltic Sea. Applied Soft Computing, 26, 180–192.