Determining the Degree of Preference Liking and Effectiveness of
DMUs over a Long Period of Time by Means of a New Approach based
on the Cross Efficiency Value Chain in Harmony with Fuzzy Arithmetic
BLERTA (KRISTO) NAZARKO
Department of Informatics and Technology,
European University of Tirana,
Kompleksi Xhura, Rruga Xhanfize Keko,
ALBANIA
Abstract: - In this paper, an alternative approach is presented for the evaluation of the likeability preference and
effectiveness of DMUs, based on the DEA and fuzzy DEA models. In the magnitudes of variable values
according to input-output levels, over the time period, some have not completely clear (fuzzy) values obtained
from perceptions and surveys. For a more realistic assessment of effectiveness and determination of the degree
of preference liking, to avoid accidental fluctuation values, and to get as close as possible to the trend of the
process's progress, dynamic analysis of smoothing of the time series is applied to the input-output value levels.
This is done according to a k-order moving average, determining the new levels of the input-output values. The
approach is applied in two phases. In the first phase, the efficiency value chain matrix is determined, applying
conventional DEA models with constant and variable returns to scale, evaluation of super efficiencies, fuzzy
efficiency, and cross-efficiency. The data and the comparison of the models are analyzed, focusing in particular
on the cross-efficiency value chain. In the second phase, fuzzy triangular numbers are composed of of the chain
of cross-efficiency values for each DMU. Then based on fuzzy arithmetic as well as the concept given by the
geometric probability model is determined and the transition matrix of the degree of preference liking, the
evaluation of the ranking is obtained according to the degree of preference liking of each DMU in relation to
other DMUs. In the paper, the contributions of the approach to the evaluation of the effectiveness and the
degree of preference liking with the relevant conclusions are highlighted.
Key-Words: - cross-efficiency, degree of preference liking, ranking, dynamic analysis, DEA models, fuzzy
triangular numbers, super efficiencies.
5HFHLYHG0D\5HYLVHG2FWREHU$FFHSWHG1RYHPEHU3XEOLVKHG'HFHPEHU
1 Introduction
The approach presented is based on real-life data,
which aims to explore the distinctive features of
each DMU assessed as best practices as well as to
identify the impact of factors on the effectiveness
and the degree of preference likeability related to
the standard of living on the basis of the prefecture
and region. In Albania, according to regions and
prefectures, there are tangible differences in internal
migration as well as in the natural rate of population
growth. For the study analysis, the method of data
envelopment analysis (DEA) is used as a very
applied and powerful method in the study and
evaluation of the effectiveness, ranking, and
evaluation of the influence of the factors on the
efficiency value. In real life, the values of the
variable quantities are not all completely
determined, where DEA also shows "weakness" if
the inputs and outputs with which the DMUs
operate do not have completely clear quantitative
values or the data are vague as they can be those
given in the field of perceptions, surveys, etc.
Therefore, in addition to the basic DEA models with
constant returns to scale and variable returns to
scale, [1], [2], many models of its extension are also
applied, as well as Fuzzy DEA models. The study
analysis is over an extended period of time (2016-
2022), so the data can be presented formatted as a
time series, where seasonal variations, random
variations (or accidental variations influenced by the
conditions in which the observations are carried out)
are encountered, etc. In order to soften the variations
in the values of variable quantities and to get as
close as possible to the trend of the period,
decomposing time series and smoothing
transformations on time series is done, as is the
series of moving averages of order k, [3]. In the
multiplier model with linear programming in the
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
393
Volume 19, 2024
DEA method, the DMUs choose their most
favorable weights, but the weights can have zero
values or very small values where it can be said that
specific inputs or outputs are ignored or
misinterpreted, and the performance of DMUs
decreasing and the power of distinguished between
them. To increase the distinguishing power in the
DEA rankings, many approaches have been applied,
such as the super-efficiency approach developed for
the first time by [4], as well as the connection of the
DEA performance with other approaches with
canonical correlation analysis such as [5]. The
approach called cross efficiency was presented for
the first time by [6] for evaluating the performance
of DMUs, but it can also be said as a likability
evaluation for management strategies. The
evaluation of efficiency of each DMU is evaluated
with its own weights, but also with the weights of
other units, which is called cross efficiency, where
each DMU is compared with every other unit in the
set of DMUs, then it is evaluated average efficiency
values for each DMUs. Applications of Cross-
efficiency can be found in many papers such as [7],
[8], [9], where in [9] the cross-efficiency evaluation
method is used for 102 DMUs (for the years 2012
and 2017), which use 4 inputs and 4 outputs, where
two of the outputs have a qualitative nature. Fuzzy
DEA is developed based on the theory of fuzzy sets.
[10] is the first to present the Fuzzy set, also [11], in
addition to the generalization of the conventional
Fuzzy set, connected it with the so-called
membership function, giving the concept of
linguistic variable. The authors [12] and [13] give
the classification of approaches applied in Fuzzy
DEA, classifying them into 6 types. [14] provides an
approach to Fuzzy DEA in a form characterized by
numbers reflected through perception, also
proposing an extension of the Fuzzy DEA model in
the relationship between DEA and linear regression.
The ranking is related to the comparison of Fuzzy
numbers. [15] makes a comparison of Fuzzy
numbers based on the concept of probability, giving
examples compared with other approaches. [16]
develops the approach of programming possibilities
with a certain level of possibility based on three
components, so that Fuzzy numbers are realistic to
represent approximations and use the concept of
possibility by comparing fuzzy numbers. [17]
developed the approach in the case of Fuzzy linear
programming and linear programming with multiple
objectives, giving a modified model for each case.
Based on different applications for the ranking of
fuzzy numbers, in the coefficient of variation of the
distance of the central point and the initial point [18]
proposes a modification of the approach based on
the distance called sign distance. [19] proposes a
new ranking function for the ranking of the real
number and the fuzzy number with an acceptance
rate and then extends it to the ranking of two fuzzy
numbers. The ranking of fuzzy numbers is
interpreted as an instrument in many application
models. To evaluate the measurement of efficiency
using the concept of the set of fuzzy numbers in the
context of DEA, [20] brings fuzzy mathematical
programming, to contribute to an optimal solution in
the evaluation of efficiency, fuzzy regression to
illustrate and types of different options that are
available. Efficiency evaluation and ranking of
DMUs with Fuzzy data, where the CCR fuzzy
model is transformed into a crisp linear
programming problem applying α-cut approach
illustrated and with numerical examples is given in
[21]. [22] presents fuzzy DEA models based on
fuzzy arithmetic formulated as a linear
programming where the fuzzy efficiency of
decision-making units can be evaluated and an
analytical approach of fuzzy ranking developed
according to fuzzy rank efficiencies for performance
evaluation. [23] proposed finding a common set of
weights in fuzzy DEA by evaluating the upper
bounds of the weights in the solution of the problem
presented in linear programming., demonstrate the
flexibility of the procedure illustrated and with
examples. In [24] a fuzzy expected value approach
is proposed for DEA analysis, in which we first
obtain the weights of the values for the inputs and
outputs. These weights are used to measure the
optimistic and pessimistic efficiency of DMUs.
Then the geometric mean is evaluated. Fuzzy
models are built based on fuzzy arithmetic and α-
level sets, determining the ranking approach for
fuzzy efficiencies. [25] provides a model of fuzzy
DEA dynamics in a study to compare discriminating
power and perceived improvement with the aim of
improving the performance of DMUs operating with
56 railways in computational time and
discriminating power. [26] presents the model in the
fuzzy context to evaluate efficiency and productivity
in an uncertain environment with different α levels,
where decision-makers can evaluate economic and
environmental factors in the selection of sustainable
suppliers with a probability distribution. [27]
presents a new approach for priorities in the process
of fuzzy analytical hierarchy, where the fuzzy nature
of the data is maintained in all the steps of the
approach, further determines the level of
consistency, gives the pairwise comparison matrix
with appropriate index with the aim of selecting a
better ventilation system. Considering the input and
output data that may be inaccurate in [28] a possible
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
394
Volume 19, 2024
approach to solving DEA models with fuzzy data is
proposed, where the fuzzy data are constructed as
trapezoidal fuzzy numbers to evaluate the technical
efficiency. To study the impact of undesirable
factors in a banking system where 12 DMUs operate
[29] analyses their effectiveness by proposing the
integration of the cross-efficiency model from DEA
and the α-cut model of fuzzy DEA. In the approach
presented in [30], the approach for an extension of
DEA is presented, where more components are
operated with variables that take the form of fuzzy
numbers for the evaluation of the fuzzy technical
efficiency and the ranking of DMUs, giving linear
programming in numerical applications. The
application of Fuzzy DEA in a two-phase process
with multiple objectives investigating the effects on
the efficiency value with the application of multiple
linear regression is seen in [31]. [32] presents an
approach based on the lexicographic language in a
linear program with many objectives for efficiency
evaluation in fuzzy efficiency models. This paper
presents an alternative approach for evaluating the
performance of DMUs along a time course
according to a given period with the evaluation of
the degree of preference based on Fuzzy arithmetic.
The evaluation of the efficiency value chain is given
by the application of DEA models, CCR efficiency,
super efficiency, Cross-efficiency value chain, and
Fuzzy efficiency where the respective rankings are
determined. The data of variable quantities are
processed in what is called statistics smoothing of
the time series. This is done as the data may not be
clearly defined (obtained from perceptions or from
surveys). The evaluation of the efficiency according
to the different models is done to provide the most
detailed performance evaluation, comparing the
models and the advantage of the application of
Cross efficiency in the composition of fuzzy
numbers and the application in the evaluation of the
degree of preference in relation to the others
models. For the evaluation of the degree of
preference based on Fuzzy arithmetic, from the
chain of cross-efficiency values evaluated in the
relevant time course, the composition of fuzzy
numbers is done first. In harmony with fuzzy
arithmetic, for two Fuzzy triangular numbers (),
where is again a triangular fuzzy number,
which can be projected geometrically as a location,
and then the geometric model of the probability of
the event is applied, enabling presentation of the
matrix of degree of preference. This alternative
approach can be applied in certain fields in the
evaluation of the degree of preference of DMUs that
operate along a given time course.
2 Methodology
The study of this work in the basic conception is
based on first evaluating the data in their dynamic
analysis, displayed during a period of time (which
can be obtained from perceptions and surveys). By
evaluating the efficiency of DMUs according to the
values obtained from data processing according to a
module k (or order k), , determined according
to the moving average (dynamic time series
analysis). The data is applied to DEA and Fuzzy
DEA models for evaluating effectiveness, ranking
DMUs, and relevant analysis by comparing the
models applied, and determining the advantages of
each model. Over an extended period of time, data
can be viewed as a time series for any variable. If
the time series is given followed by consecutive
values X1, X2, …, XN, we determine the “series” of
values according to the module of the moving
averages. The series of values according to the
model is given by the countable sums, [33],
[34]:
:
,
, …
,
.
These values, defined as input-output levels, are
applied for a more realistic assessment and
examination of the periodic trend of instantaneous
changes and the evaluation of the effectiveness of
each DMU according to the relevant DEA models.
By successively following the chain of efficiency
values with the data according to the module,
the matrix of the chain with efficiency values is also
built.
Efficiency values are calculated by applying the
DEA model input oriented according to constant
and variable returns to scale (CRS and VRS), [35]:
= min (1)
s.t: -
0 =1,2,…,m,
= min (2)
s.t
=1,2,…,m,
,
0
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
395
Volume 19, 2024
The matrix is constructed with the chain of
efficiency values as below, also giving the indicator
of the change in the efficiency value (Ind. EC).
=
,
where, Ind. EC =
To increase the distinguishing power in the
ranking of DMUs (where several DMUs with the
efficiency 1), the super-efficiency
evaluation model is applied as follows [4], [35]:
min
st:
i=1,2,…,m;
(3)
, r =1,2..,s;
, j
Based on the conclusions of the efficiency
values (CRS and VRS), the scale efficiency is also
determined, which also enables the classification of
inefficiencies, [36]. For each DMU, the impact of
variable factors on efficiency values is also
determined using the formula:
W() =
100% ( is the
efficiency value according to the I-th input.)
Following the study, to determine the degree of
preference likeability based on efficiency values,
triangular fuzzy numbers are composed (from
matrices of efficiency values of one model).
Fuzzy background:
The definition of fuzzy set theory was first proposed
by [10]. Based on this concept, the field of
applications has been expanded in the presentation
of solutions to problems that use it in research
operations and other areas from the practical life of
the real world, where the data are not completely
clearly defined. Fuzzy sets are related to a
characteristic function called the membership
function. In accordance with the definitions given
by [37], [38] are given:
Definition 1: A Fuzzy set given over a collection
of objects X is defined by the set of ordered pairs
= {(x, : xX}, where is the
membership function, of the membership degree
value of each element x and that : X [0,1].
Definition 2: The set of elements of x (collection)
such that the value of the membership function is
equal to 1,=1, represents Core ( ), core of
Fuzzy set.
Mathematically, this can be given by the
equation core(= {x:=1}.
Definition 3: (α-cut) (Figure 1)
Let it be the geometric projection of a
trapezoidal number Fuzzy = (, where
are all real numbers (
).
Fig. 1: core, –cut, and support of fuzzy set, [38]
The α-cut set of a Fuzzy number (trapezoidal)
is the set given and denoted ={x:
}. Let us express the membership function,
=
where
and
are membership functions
of the left side and the right side. If according to (α-
cut) we have = then we say that:
x =
for x [, = 1.
In Figure 1, where the geometric projection of
the Fuzzy number is given, core (), cut, and
support of are presented respectively. Support of
Fuzzy set is presented when α-cut we have α=0, 0-
cut. For a Fuzzy triangular number = (
we have the following :
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
396
Volume 19, 2024
=
Let two Fuzzy numbers be given: =
( and = ( and k a
scalar [38], [37], [39]:
a) Addition of two fuzzy numbers: = =
().
b) Subtraction of two fuzzy numbers: = =
().
c) Multiplication of a fuzzy number by a scalar
number:= =
• Positive Fuzzy Triangular Numbers and Negative
Fuzzy Triangular Numbers
A triangular fuzzy number given =
( is classified as positive if
(i=1,2,3) and classified as negative if
(i=1,2,3).
They are classified as partial negative if
.
If we have n DMUs, where the levels of inputs
and outputs can also be characterized by fuzzy
triangular numbers (= (
,
,
) and =
(
,
,
).The values of fuzzy efficiencies
() [22] are calculated from the models
(4), (5) and (6).
max =
st:
= 1 (4)
-
0
, r ,, i.
max =
st:
= 1 (5)
-
0
, r ,, i.
max =
st:
= 1 (6)
-
0
, r ,, i.
To evaluate the ranking of DMUs, the
geometric mean is evaluated =
[40].
• Cross-efficiency evaluation
In addition to the evaluation of efficiencies
according to the above models, the cross efficiency
evaluation model is also applied, which has a better
distinguishing power for ranking DMUs, [35].
In conventional DEA models, DMUs have the
nature of self-evaluation in the selection of weights
for each input and output, where we can have
several efficient evaluated DMUs, where Ef=1.
The Cross efficiency model is the approach
presented by [6] and [41]. The efficiency values of
each DMUs with the cross efficiency model are
evaluated not only with their own weights, but also
with the input -output weights of other units. The
average representing the Cross efficiency result for
a DMUj (j=1, 2,...,n) is given:
=
,
where this average is the average of the values
according to the column presented in the Cross
efficiency matrix, [41]. From the cross efficiency
value chain, the corresponding efficiency value
chain matrix is formatted, where the harmonic cross
efficiency for the given period and the
corresponding rankings are also evaluated. From the
cross efficiency value chain matrix for each DMU,
the triangular fuzzy numbers , are composed,
where (, where this matrix fully
enables this composition, which is advantageous
over other models.
After determining the triangular fuzzy numbers
from the cross efficiency values, based on fuzzy
arithmetic and the geometric model of the concept
of event probability, the approach for determining
the degree of preference likeability is then applied.
In the following figures, the cases of geometric
projections of triangular fuzzy numbers are given
when their comparison is required. If two triangular
fuzzy numbers = () and = (
according to fuzzy arithmetic and fuzzy triangular
number =( = ()
where for each of them their geometric projections
can be given in a coordinate plane. For each case, it
can be judged according to the concept of the
geometric model of the probability of event A, the
probability of this event is estimated. Case 1: In the
Figure 2, the geometric projection of the two fuzzy
numbers = () = (5, 7, 9.5) and =
( = (1, 2, 3.5) is given. The fuzzy
triangular number =( = (1.5, 5, 8.5), where
it seems that the fuzzy triangular number is
classified as positive, since each of its components
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
397
Volume 19, 2024
is greater than zero, so it can be said that in this case
P(A) = () =1. From the projections, it is noted
that the condition () must be satisfied.
Fig. 2: Geometric projections of fuzzy numbers, =
() = (5, 7, 9.5) and = ( = (1, 2,
3.5)
Case 2: (Figure 3) If two triangular fuzzy numbers
= (2, 4, 5) and = (5.5, 6, 7) are given, as well as
=( = (-5, -2, -0.5), where it can be said that
the triangular fuzzy number =( is classified
as negative, so in this case it is said that P(A) =
() = 0. From the projections it is noted that
the condition must be fulfilled.
Fig. 3: Geometric projections of fuzzy numbers, =
(2, 4, 5) and = (5.5, 6, 7)
Case 3: In the Figure 4 and Figure 5 if fuzzy
triangular numbers = (1, 4, 6), = (2, 5, 7) and
= ( = (-6, -1, 4) are given. This number is
partial negative triangular fuzzy number, so based
on the concept of the geometric model of the
probability of the event P(A) = P() =
, where S(p) is the area of the event A and
= S() is the area containing all the
elementary events of a zone Z. Ω = A dhe A
= . In the case of the Figure 5.
Fig. 4: Geometric projections of fuzzy numbers, =
(1, 4, 6), = (2, 5, 7)
Fig. 5: Geometric projections of numbers, (-
) = (-6, -1, 4)
S(p) =
dx =
.
S(Ω) = S( =
[ (,
so P()=
,
from the presented case it is said that the condition
() ( is fulfilled. Case 4: (Figure 6
and Figure 7) Where = (1.5, 4.5, 5.5), = (0.5,
2.5, 3.5) and = (-2, 2, 5).
Fig. 6: Geometric projections of fuzzy numbers, =
(1.5, 4.5, 5.5), number (-) = (-2, 2, 5)
Fig. 7: Geometric projections of fuzzy = (0.5, 2.5,
3.5)
In this case, it can be written P() = 1-
. It is calculated
=
.
P()= 1-
.
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
398
Volume 19, 2024
From the presented case it is said that the condition
() ( is fulfilled.
Based on the calculation of the probabilities
according to the cases presented above, the matrix
of the degree of liking and preference is formed as
follows
=
.
Statistical tests are also applied in the paper.
3 Numerical Application
The aim of the numerical application is to evaluate
the effectiveness and determine the degree of
likability preference in the performance presented
by 15 DMUs (the 15 DMUs are the administrative
divisions in the prefectures and regions in Albania).
The study covers the 7-year period (2016-2022).
The prefectures and regions between them show
differences in their multifaceted development of
production and standard of living. DMUs operate
with three inputs and two outputs. The data are
taken from the Statistical Yearbooks (INSTAT)
(2017-2023) and the Labour Market (2016-2023)
[42], [43]. Both outputs carry values as variables not
fully determined (surveys): 1. Distribution of
households according to the area used for housing
(over 90 square meters, in percentage), 2.
Ownership of long-term devices with multiple
functions in households according to prefectures (in
percentage). The three inputs are related to
employment in the labour market according to the
three main sectors of the economy in percentage
(agriculture, industry and construction, services) for
each prefecture and region. The study follows the
following course:
- smoothing of the values that present as a time
series each input and output according to the
movement of the average of the order k (treated in
the methodology)
Table 1 (Appendix) shows the values of the
overall average data for each variable size.
The application of models is done according to the
following steps:
Step 1. Application of linear programming models
((1), (2), (3)) for evaluation of CRS, VRS, super -
efficiencies for efficient units evaluated by the CRS
efficiency model. This evaluation is done according
to each grouping with Eff-(3) (i=1,2,3,4,5).
Classifications of the source of inefficiencies are
given by comparing scale efficiency with VRS
efficiency.
Step 2: Evaluation of Cross efficiency according to
each grouping Eff-(3) (i=1,2,3,4,5), calculation
and evaluation of fuzzy technical efficiency.
The impact of variable factors on the cross-eff value
is calculated, at the beginning of the period (Eff -
(3)) and at the end of the period (Eff -(3)),
(multiple linear regression) where efficiency values
are considered as dependent variable and the
respective output values as independent variables.
The summary results are given in Table 4
(Appendix).
With significance level = 0.05, = 3.89
where from the Table 3 (Appendix), F
for both cases. Then the hypothesis = 0
is rejected. Table 4 (Appendix) shows that at the
beginning of the period the first output has the
greatest impact (
) while at
the end of the period the second output has the
greatest impact ((
The results obtained from the second step are
given in Table 5 (Appendix).
From the summary rankings obtained from the
three models, the overall ranking is estimated. Table
6 (Appendix) presents the overall ranking estimate
as well as the triangular fuzzy numbers composed of
the cross efficiency value matrix.
Step 3: Determining the degree of preference liking
based on fuzzy arithmetic and the concept of the
geometric probability model (based on the
methodology). The composition of fuzzy numbers is
given according to the matrix of cross-efficiency
values given in Appendix in Table 5 and (Appendix)
6. The matrix of cross efficiency values is used,
according to the fuzzy efficiency values
( there are DMUs that have the fuzzy
efficiency values =1 as D9 and D4
(=1).
From Table 7 (Appendix), it is noted the
comparative preference between the decision-
making units.
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
399
Volume 19, 2024
where
shows that has a better
performance than to the extent to 74.87%, the
same can be said for the others.
In the summary of the results obtained according to
the rankings, it is noted that the Spearman
coefficient () shows these values:
1) The rank correlation between: the cross
efficiency (harmonic) ranking and the
ranking of the degree of preference, =
0.98.
2) The rank correlation between: CRS
efficiency ranking (overall efficiency) and
preference degree ranking, = 0.95
3) The rank correlation between: CRS (overall
efficiency) efficiency ranking and Cross
(harmonic) efficiency ranking, = 0.90
4) The rank correlation between: ranking of
the fuzzy geometric mean and overall
efficiency, = 0.87
5) The rank correlation between: overall
ranking (Table 6, Appendix) and degree of
liking preference ranking (Table 7,
Appendix), = 0.98
For each of the above cases, the hypothesis test
is taken: : = 0 dhe 0. For n=15 and
a significant level of = 0.005, the critical value is
0.654. This shows that is rejected.
Even applying the statistic_t, where t =
, ku =
and r = . For the level of = 0.05,
= = 2.160. For the above five
points, the calculated t-values are =17.756; =
10.97; = 7.44; = 6.3025 and = 17.756, so
is rejected.
The values of the Spearman rank correlation
coefficient and hypothesis testing indicate strong
correlations between the rankings.
In conclusion of the results obtained, it can be
said that certain DEA models can be used to assess
the effectiveness of DMUs with appropriate
objectives. From the above results, the following
DMUs are found to be best practices:
,, . The general feature for them is the
impact that employment represents in the service
sector and in the industry and construction, being
higher than other DMUs (Table 2, Table 3 and
Table 4 in Appendix). In the classification of
inefficiencies (Table 2, Appendix), managerial
inefficiency is most evident. While the weakest
practices are the DMUs that have the highest
percentage of employment in agriculture.
While assessing the degree of preference liking,
the most advantageous approach is the application
of cross efficiency in harmony with Fuzzy
arithmetic and the geometric model of probability.
From the results obtained with the data in the above
tables, the performance of each DMU is determined
according to the objectives, the degree of
preference, and the analysis judgments.
4 Conclusion
In this work, the evaluation of the performance of
DMUs was dealt with by determining the degree of
liking preference and effectiveness (efficiency
value) of each DMU according to the input-output
levels with the processed data as well as
determining the respective rankings. The levels of
quantities of input-output values where DMUs
operate, where some of them do have not
completely clear (fuzzy) values, according to the
principle of dynamic analysis of a time series,
smoothing of the time series is applied by
determining the input-output levels in processed
values, which are applied to the respective models.
Conventional DEA models, cross-efficiency model,
super efficiency models, Fuzzy efficiency
evaluation models are applied in this paper. The
alternative approach was applied according to a
two-phase process, where the degree of liking
preference was determined based on the cross-
efficiency value chain in harmony with fuzzy
arithmetic. We emphasize that in harmony with the
fuzzy arithmetic, the matrix of the cross efficiency
value chain was used, because with the fuzzy
efficiency values () it may happen that we
can have DMU where these values are equal or two
of them are equal. Thus, two of the DMUs
(specifically D4 and D9) do not enable the
composition of the fuzzy number (in the order
), so the model of the application of the
value chain of cross efficiency enables the
composition of triangular fuzzy numbers, so it can
be said that it is more advantageous to apply it in
harmony with fuzzy arithmetic. In summary, the
contribution of the application of this approach can
be said to be in fulfilling these objectives:
1. We better evaluate the meaningfulness of
uncertain (fuzzy) data for their application in
DEA and DEA Fuzzy models. Using several
models provides a more realistic and
informative assessment of effectiveness.
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
400
Volume 19, 2024
2. An alternative method is reflected as an
approach of a two-phase process. The first
phase determines efficiency value matrices
according to models that assess the efficiency of
DMUs. In the second phase, fuzzy numbers are
composed of the cross-efficiency value matrix,
and based on fuzzy arithmetic and the concept
of the geometric probability model, the degree
of preference likeability and preference ranking
are determined.
The first phase determines the matrix of the
value of the cross efficiency along an extended
time course and the second phase, based on
fuzzy arithmetic and the concept of the
geometric model of probability determines the
degree of likability preference and preference
ranking.
3. The evaluation of the application of this
approach shows the advantages of the
application in the evaluation of the effectiveness
and the determination of the degree of likability
preference over a long period of time in relation
to the conventional DEA models, which also
present limitations. The review and analysis of
data according to each model is also selected
based on the relevant objectives that can be set.
References:
[1] A.Charnes,W.W.Cooper, E.Rhodes,
“Measuring the Efficiency of Decision
Making Units” (1978), European Journal of
Operational Research, vol. 2, Issue 6, 429-
444, https://doi.org/10.1016/0377-
2217(78)90138-8.
[2] Banker, R.D., Charnes, A., Cooper, W.W.
(1984) “Some models for estimating technical
and scale inefficiencies in data envelopment
analysis”, Management Science, Volume 30,
No.9, 1078-1092, DOI:
https://doi.org/10.1287/mnsc.30.9.1078.
[3] Terence C. Mills. (2019). Applied Time Series
Analysis. A Practical Guide to Modeling and
Forecasting, Candice Janco. ISBN: 978-0-12-
813117-6.
[4] Andersen, P. and Petersen, N.C. (1993) A
Procedure for Ranking Efficient Units in Data
Envelopment Analysis. Management Science,
39, 1261-1264.
http://dx.doi.org/10.1287/mnsc.39.10.1261.
[5] Lea Friedman, Zilla Sinuany-Stern. (1997).
Scaling units via the canonical correlation
analysis in the DEA context, European
Journal of Operational Research, Vol. 100,
Issue 3, pp.629-637,
https://doi.org/10.1016/S0377-
2217(97)84108-2.
[6] Sexton,T.R., Silkman, R. H., Hogan,A.J.
(1986). Data envelopment analysis:Critique
and extensions. In R.H Silkman (Ed),
Measuring efficiency: An assessment of data
envelopment analysis. Vol 32, pp 73-105, San
Francisco: Jossey-Bass.
[7] Jie, Wu., Liang, Liang., Yingchun, Zha.
(2009). Preference voting and ranking using
DEA game cross efficiency model.
(Operations Research for Performance
Evaluation). Journal of The Operations
Research Society of Japan, 52(2):105-111.
doi: 10.15807/JORSJ.52.105.
[8] Cook, W.D., & Zhu,J. (2014). DEA Cobb-
Douglas frontier and cross efficiency. Journal
of operational research society, 65 (2), 265-
268.
[9] Bert M. Balk, M.B.M. (René) De Koster,
Christian Kaps, José L. Zofío, (2021). An
evaluation of cross-efficiency methods: With
an application to warehouse performance,
Applied Mathematics and Computation, Vol.
406, 126261, pp 1-14,
https://doi.org/10.1016/j.amc.2021.126261.
[10] L. A. Zadeh. (1965). Fuzzy sets. Information
and Control. Vol. 8, Issue 3, pp.338-353,
https://doi.org/10.1016/S0019-
9958(65)90241-X.
[11] L.A.Zadeh (1975). The concept of linguistic
variable and its application to approximate
reasoning-II, Information Sciences, Vol. 8,
Issue 4, pp. 301-357,
https://doi.org/10.1016/0020-0255(75)90046-
8.
[12] Ebrahimnejad, Ali & Amani, Naser, (2020).
Fuzzy data envelopment analysis in the
presence of undesirable outputs with ideal
points, Complex & Intelligent Systems, vol. 7.
pp.1-22, http://dx.doi.org/10.1007/s40747-
020-00211-x.
[13] Izadikhah, Mohammad & Khoshroo, Alireza.
(2017) “Energy management in crop
production using a novel Fuzzy Data
Envelopment Analysis model”, RAIRO -
Operations Research. 52, 2018, 595-617,
https://doi.org/10.1051/ro/2017082.
[14] Peijun Guo, Hideo Tanaka, (2001).Fuzzy
DEA: a perceptual evaluation method. Fuzzy
Sets and Systems, Vol. 119, Issue 1, pp.149-
160, https://doi.org/10.1016/S0165-
0114(99)00106-2.
[15] E.S. Lee, R.-J. Li. (1988). Comparison of
fuzzy numbers based on the probability
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
401
Volume 19, 2024
measure of fuzzy event. Computers &
Mathematics with Applications, Vol. 15, Issue
10,
pp. 887-896, https://doi.org/10.1016/0898-
1221(88)90124-1.
[16] Devendra S. Negi, E. Stanley Lee. (1993).
Possibility programming by the comparison of
Fuzzy numbers. Computers & Mathematics
with Applications. Vol. 25, Issue 9. pp. 43-50.
https://doi.org/10.1016/0898-1221(93)90131-
E.
[17] M.G. Iskander. (2002). Comparison of Fuzzy
numbers using possibility programming:
Comments and new concepts. Computers &
Mathematics with Applications. Vol. 43,
Issues 6–7, pp.833-840.
https://doi.org/10.1016/S0898-
1221(01)00324-8.
[18] S. Abbasbandy, B. Asady. (2006). Ranking of
fuzzy numbers by sign distance. Information
Sciences, Vol. 176, Issue 16, pp.2405-2416,
https://doi.org/10.1016/j.ins.2005.03.013.
[19] Yaghobi, Mohamad & Rabbani, Mohsen &
Adabitabar firozja, M. & Vahidi, Javad.
(2014). Comparison of fuzzy numbers with
ranking fuzzy and real number. Journal of
Mathematics and Computer Science. 12.
pp.65-72. doi:10.22436/jmcs.012.01.06.
[20] Jati K. Sengupta. (1992). A fuzzy systems
approach in data envelopment analysis.
Computers & Mathematics with Applications.
Vol. 24, Issues 8–9, pp. 259-266,
https://doi.org/10.1016/0898-1221(92)90203-
T.
[21] Saati, S.M., Memariani, A. & Jahanshahloo,
G.R. (2002). Efficiency analysis and ranking
of DMUs with fuzzy data Fuzzy. Optimization
and Decision Making, 1, (3) 255–267,
https://doi.org/10.1023/A:1019648512614.
[22] Ying-Ming Wang, Ying Luo, Liang Liang.
(2009) Fuzzy data envelopment analysis
based upon fuzzy arithmetic with an
application to performance assessment of
manufacturing enterprises. Expert Systems
with Applications, Vol. 36, Issue 3, Part 1,
pp.5205-5211,
https://doi.org/10.1016/j.eswa.2008.06.102.
[23] S. Saati, A. Memariani, (2005) Reducing
weight flexibility in fuzzy DEA, Applied
Mathematics and Computation, Vol. 161,
Issue 2, pp.611-622,
https://doi.org/10.1016/j.amc.2003.12.052.
[24] Ying-Ming Wang, Kwai-Sang Chin. (2011).
Fuzzy data envelopment analysis: a fuzzy
expected value approach. Expert Systems with
Applications. Vol. 38, Issue 9, pp.11678-
11685,
https://doi.org/10.1016/j.eswa.2011.03.049.
[25] Jafarian-Moghaddam, A.R., Ghoseiri, K.
(2012). Multi-objective data envelopment
analysis model in fuzzy dynamic environment
with missing values. The International
Journal of Advanced Manufacturing
Technology. Vol. 61. pp.771–785.
https://doi.org/10.1007/s00170-011-3730-7.
[26] Majid Azadi, Mostafa Jafarian, Reza
Farzipoor Saen, Seyed Mostafa
Mirhedayatian. (2015). A new fuzzy DEA
model for evaluation of efficiency and
effectiveness of suppliers in sustainable
supply chain management context. Computers
& Operations Research, Vol. 54, pp.274-285,
https://doi.org/10.1016/j.cor.2014.03.002.
[27] Mirhedayatian, Seyed & jafarian jelodar,
Mostafa & Adnani, Siamak & Akbarnejad,
Mohammad & Farzipoor Saen, Reza. (2013).
A new approach for prioritization in fuzzy
AHP with an application for selecting the best
tunnel ventilation system. The International
Journal of Advanced Manufacturing
Technology. 68. (9-12). doi: 10.1007/s00170-
013-4856-6.
[28] Nedeljković, Ranko & Drenovac, Dragana.
(2012). Efficiency Measurement of Delivery
Post Offices Using Fuzzy Data Envelopment
Analysis (Possibility Approach). International
Journal for Traffic and Transport
Engineering. Vol. 2, 22-29, [Online].
https://api.semanticscholar.org/CorpusID:119
72655 (Accessed Date: July 11, 2024).
[29] Jolly Puri, Shiv Prasad Yadav. (2014).A fuzzy
DEA model with undesirable fuzzy outputs
and its application to the banking sector in
India. Expert Systems with Applications. Vol.
41, Issue 14, pp.6419-6432,
https://doi.org/10.1016/j.eswa.2014.04.013.
[30] Puri, J., Yadav, S.P. (2015) A fully fuzzy
approach to DEA and multi-component DEA
for measuring fuzzy technical efficiencies in
the presence of undesirable
outputs. International Journal of System
Assurance Engineering and
Management. Vol. 6, pp.268–285
https://doi.org/10.1007/s13198-015-0348-4.
[31] Boubaker, Sabri & Do, Duc & Hammami,
Helmi & Ly, Kim. (2022). The role of bank
affiliation in bank efficiency: a fuzzy multi-
objective data envelopment analysis approach.
Annals of Operations Research. 311 (2): 1-29.
doi: 10.1007/s10479-020-03817-z.
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
402
Volume 19, 2024
[32] Adel Hatami-Marbini, Ali Ebrahimnejad,
Sebastián Lozano.(2017).Fuzzy efficiency
measures in data envelopment analysis using
lexicographic multiobjective approach.
Computers & Industrial Engineering, Vol.
105, pp.362-376,
https://doi.org/10.1016/j.cie.2017.01.009.
[33] Yadolah Dodge. (2008). The Concise
Encyclopedia of Statistics. Springer New
York, NY. https://doi.org/10.1007/978-0-387-
32833-1.
[34] Douglas C. Montgomery, Cheryl l. Jennings,
Murat Kulahci. (2008). Introduction to time
series analysis and forecasting. John Wiley &
Sons.
[35] Zhu, Joe. (2014). Quantitative models for
performance evaluation and benchmarking.
Data envelopment analysis with spreadsheets.
3rd ed. International Series in Operations
Research and Management Science, Springer,
doi: 10.1007/978-3-319-06647-9.
[36] William W. Cooper, Lawrence M. Seiford,
Kaoru Tone. (2002). Data Envelopment
Analysis. A Comprehensive Text with Models,
Applications, References and DEA-Solver
Software, Kluwer Academic Publishers, New
York, Boston, Dordrecht, London,Moscow,
ISBN: 0-7923-8693-0.
[37] Zimmermann, H.J.,” Fuzzy Set Theory — and
Its Applications”, Springer Netherlands,
ISBN: 9789401579490, 2013, [Online].
https://books.google.al/books?id=GDfpCAA
AQBAJ (Accessed Date: August 10, 2024).
[38] Asesh Kumar Mukherjee, Kamal Hossain
Gazi, Soheil Salahshour, Arijit Ghosh, Sankar
Prasad Mondal.(2013)A Brief Analysis and
Interpretation on Arithmetic Operations of
Fuzzy Numbers, Results in Control and
Optimization, Vol. 13,100312,
https://doi.org/10.1016/j.rico.2023.100312.
[39] Dubois, D., & Prade, H. (1978). Operations
on fuzzy numbers. International Journal of
Systems Science, 9(6), 613–626.
https://doi.org/10.1080/00207727808941724.
[40] Blerta (Kristo) Nazarko (2024). Ranking of
DMUs and Evaluation of the Impact of
Variable Factors Using Fuzzy DEA. Al
Farabi 13th International Scientific Research
and Innovation Congress, IKSAD, Turkey,
ISBN: 978-625-367-928-6
[41] Doyle, J., & Green, R. (1994). Efficiency and
Cross-Efficiency in DEA: Derivations,
Meanings and Uses. The Journal of the
Operational Research Society, 45(5), 567–578.
https://doi.org/10.2307/2584392.
[42] Labor Market (Tregu i Punës) from 2016 to
2023, INSTAT, [Online].
https://www.instat.gov.al/al/publikime/librat/
(Accessed Date: May 07, 2024).
[43] Regional Statistical Yearbook (Vjetari
Statistikor Rajonal) from 2017 to 2023,
INSTAT, [Online]
https://www.instat.gov.al/al/publikime/librat/
(Accessed Date: April 6, 2024).
Contribution of Individual Authors to the
Creation of a Scientific Article (Ghostwriting
Policy)
The author contributed in the present research, at all
stages from the formulation of the problem to the
final findings and solution.
Sources of Funding for Research Presented in a
Scientific Article or Scientific Article Itself
No funding was received for conducting this study.
Conflict of Interest
The author has no conflicts of interest to declare.
Creative Commons Attribution License 4.0
(Attribution 4.0 International, CC BY 4.0)
This article is published under the terms of the
Creative Commons Attribution License 4.0
https://creativecommons.org/licenses/by/4.0/deed.en
_US
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
403
Volume 19, 2024
APPENDIX
Table 1. Average data values (for 15 DMUs) over time
I/O
Average data for 15 DMUs (%)
The data by smoothing of time series (%)
46.3
43.5
43.5
43.4
43.9
42.0
41.2
44.4
43.5
43.6
43.1
42.4
18.4
18.5
18.2
18.6
18.8
20.2
19.6
18.4
18.4
18.5
19.2
19.5
35.3
38.1
38.3
38.2
37.3
37.8
39.1
37.2
38.2
37.9
37.8
38.1
38.7
39.7
38.1
36.6
38.6
37.8
42.3
38.8
38.1
37.8
37.7
39.6
30.3
42.5
50.9
60.4
64.8
68.6
75.7
41.2
51.3
58.7
64.6
69.7
Table 2. Values of efficiency, super -efficiency, rankings and classification of sources of inefficiencies
according to each grouping
Note: Eff- values of efficiency and super efficiency; C-- Classification of inefficiencies [38] (a-management inefficiency, 1,
1 and SE , SE-scale efficiency; b-scale inefficiency, 1, =1 dhe SE ; c-managerial inefficiency
and scale inefficiency, both together,1, 1 dhe SE; eff.- Efficient DMU); Ind. EC =
(k indicates
the number of grouping (in the case of the study k= 5) according to the module of moving averages of order k);
Table 3. The impact of each input on the efficiency value ()
DMU
Eff -(3)
Eff -(3)
Eff -(3)
Eff -(3)
Eff -(3)
Ind.
CE
Overall
Rank
Eff.
C-
.
Rank
Eff.
C-
.
Rank
Eff.
C-
.
Rank
Eff.
C-
.
Rank
Eff.
C-
.
Rank
D1
0.906
b
9
0.853
a
10
0.902
a
9
0.954
a
8
0.967
b
9
1.017
8
D2
0.912
a
8
0.751
a
15
0.854
b
13
0.849
b
14
0.821
b
15
0.981
13
D3
1.122
eff
3
1.029
eff
3
1.003
eff
5
1.001
eff
5
1.135
eff
3
1.000
3
D4
0.982
b
6
1.015
eff
4
1.152
eff
3
1.247
eff
2
1.351
eff
1
1.004
6
D5
1.178
eff
2
1.237
eff
2
1.257
eff
2
1.172
eff
3
1.053
eff
4
1.000
2
D6
0.846
a
10
0.860
a
9
0.869
a
11
0.891
a
12
0.962
b
10
1.033
10
D7
1.001
eff
4
1.000
eff
5
1.000
eff
6
1.000
eff
6
1.001
eff
5
1.000
4
D8
1.797
eff
1
1.654
eff
1
1.358
eff
1
1.386
eff
1
1.223
eff
2
1.000
1
D9
1.000
eff
5
1.000
eff
6
1.000
eff
7
1.000
eff
7
1.000
eff
6
1.000
5
D10
0.758
a
14
0.810
a
12
0.889
a
10
0.910
a
11
0.883
a
11
1.040
11
D11
0.758
b
13
0.767
b
14
0.790
b
15
0.786
b
15
0.875
b
12
1.037
15
D12
0.827
b
11
0.877
a
8
0.930
a
8
0.941
b
9
0.986
b
7
1.045
9
D13
0.947
b
7
0.997
b
7
1.080
eff
4
1.032
eff
4
0.978
b
8
1.008
7
D14
0.657
a
15
0.795
a
13
0.863
a
12
0.889
a
13
0.861
a
14
1.074
14
D15
0.791
a
12
0.843
b
11
0.840
b
14
0.915
a
10
0.873
a
13
1.026
12
W()
(3)
(3)
(3)
(3)
(3)
Average
Parameter estimation
=
W()
12.08 %
12.50%
14.59%
14.26%
14.87%
13.66%
= 0.3147
W()
39.83%
37.07%
35.87%
36.67%
33.29%
36.55%
= 1.9442
W()
48.09%
50.43%
49.54%
49.07%
51.84%
49.79%
= 1.3158
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
404
Volume 19, 2024
Table 4. Statistical results
Applied model
Multiple R
R square
F
Significance F
(3)
0.97
0.94
102.48
2.86367E-08
0.4856
0.1968
(3)
0.94
0.89
49.78
1.54903E-06
0.2454
0.3338
Table 5. Cross efficiency values, rankings according to each grouping, geometric mean efficiency (
fuzzy, ranking
DMU
Cross efficiency
Fuzzy DEA
Eff -(3)
Eff -(3)
Eff -(3)
Eff -(3)
Eff -(3)
Cross
H-Eff
Rank
Cross
Eff.
Rank
Cross
Eff.
Rank
Cross
Eff.
Rank
Cross
Eff.
Rank
Cross
Eff.
Rank
GM =
Rank
of
GM
D1
0.814
7
0.807
8
0.858
6
0.863
4
0.855
3
0.839
5
0.639
8
D2
0.769
8
0.720
13
0.770
14
0.720
14
0.672
15
0.728
13
0.568
13
D3
0.902
3
0.873
4
0.893
4
0.964
1
0.956
1
0.916
2
0.879
2
D4
0.677
13
0.675
15
0.812
11
0.851
6
0.826
7
0.760
12
0.689
6
D5
0.901
4
0.963
2
0.968
1
0.956
2
0.911
2
0.939
1
0.739
5
D6
0.756
9
0.750
12
0.814
10
0.785
12
0.811
8
0.782
9
0.597
10
D7
0.902
2
0.872
5
0.851
7
0.796
9
0.810
9
0.844
4
0.805
3
D8
0.946
1
0.996
1
0.957
2
0.821
7
0.842
4
0.907
3
0.766
4
D9
0.885
5
0.829
6
0.819
9
0.816
8
0.836
5
0.837
6
1.000
1
D10
0.693
12
0.792
9
0.838
8
0.787
11
0.751
11
0.769
11
0.587
11
D11
0.661
14
0.753
11
0.752
15
0.644
15
0.681
14
0.695
15
0.532
15
D12
0.716
10
0.826
7
0.880
5
0.790
10
0.832
6
0.805
8
0.647
7
D13
0.814
6
0.884
3
0.915
3
0.890
3
0.706
12
0.834
7
0.583
12
D14
0.598
15
0.688
14
0.776
13
0.766
13
0.706
13
0.701
14
0.538
14
D15
0.700
11
0.778
10
0.800
12
0.856
5
0.761
10
0.776
10
0.628
9
Note: Cross H-Eff- cross harmonic efficiency
Table 6. Overall ranking, composition of Fuzzy numbers from the cross efficiency value matrix
DMU
Ranking
Composition of fuzzy triangular
numbers
(Cross Eff.)
Weak
GM=
Overall
D1
8
6.83990
6
(0.807, 0.855, 0.863)
D2
13
12.99999
13
(0.672, 0.720, 0.770)
D3
3
2.28943
2
(0.873, 0.902, 0.964)
D4
12
7.55953
7
(0.675, 0.812, 0.851)
D5
5
2.15443
1
(0.901, 0.956, 0.968)
D6
10
9.65489
10
(0.750, 0.785, 0.814)
D7
4
3.63424
5
(0.796, 0.851, 0.902)
D8
4
2.28943
3
(0.821, 0.946, 0.996)
D9
6
3.04461
4
(0.816, 0.829, 0.885)
D10
11
11
12
(0.693, 0.787, 0.838)
D11
15
15
15
(0.644, 0.681, 0.753)
D12
9
7.95811
8
(0.716, 0.826, 0.880)
D13
12
8.37772
9
(0.706, 0.884, 0.915)
D14
14
14
14
(0.598, 0.706, 0.776)
D15
12
10.25986
11
(0.700, 0.778, 0.856)
Note: Weak - The weakest ranking value; GM geometric mean of rankings according to the three models applied; Overall
positioning in the overall ranking on which the effectiveness of the DMUs is judged during the period.
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
405
Volume 19, 2024
Table 7. The matrix of the degree of liking preference from Cross efficiencies and rankings according to the
degree of liking preference
D1 D2 D3 D4 D5 D6 D7 D8 D9 D10 D11 D12 D13 D14 D15 Rank
D1 …. 1.00000 0.00000 0.90349 0.00000 0.99576 0.43641 0.05863 0.52931 0.95309 1.00000 0.76358 0.50239 1.00000 0.91174 5
D2 0.00000 …0.00000 0.17571 0.00000 0.02902 0.00000 0.00000 0.00000 0.16843 0.73860 0.06836 0.05861 0.66910 0.14987 13
D3 1.00000 1.00000 …1.00000 0.21513 1.00000 0.94668 0.40988 0.98877 1.00000 1.00000 0.99781 0.90268 1.00000 1.00000 2
D4 0.09651 0.82429 0.00000 …0.00000 0.51666 0.11583 0.01619 0.09644 0.56104 0.89845 0.36096 0.25362 0.86040 0.54185 9
D5 1.00000 1.00000 0.78487 1.00000 …1.00000 0.99996 0.64191 1.00000 1.00000 1.00000 1.00000 0.99197 1.00000 1.00000 1
D6 0.00424 0.97098 0.00000 0.48334 0.00000 …0.02250 0.00000 0.00000 0.57056 0.99952 0.30192 0.20680 0.97244 0.54773 10
D7 0.56359 1.00000 0.05332 0.88417 0.00004 0.97750 …0.13275 0.58944 0.93371 1.00000 0.76057 0.53448 1.00000 0.89683 4
D8 0.94137 1.00000 0.59012 0.98381 0.35809 1.00000 0.86725 …0.90592 0.99487 1.00000 0.94258 0.85323 1.00000 0.98191 3
D9 0.47069 1.00000 0.01123 0.90356 0.00000 1.00000 0.41056 0.09408 …0.96615 1.00000 0.74076 0.49592 1.00000 0.92373 6
D10 0.04691 0.83157 0.00000 0.43896 0.00000 0.42944 0.06629 0.00513 0.03385 …0.91598 0.29834 0.21554 0.87004 0.48606 12
D11 0.00000 0.26140 0.00000 0.10155 0.00000 0.00048 0.00000 0.00000 0.00000 0.08402 …0.02662 0.02774 0.46402 0.06971 15
D12 0.23642 0.93164 0.00219 0.63904 0.00000 0.69808 0.23943 0.05742 0.25924 0.70166 0.97338 …0.35098 0.94169 0.67467 8
D13 0.49761 0.94139 0.09732 0.74638 0.00803 0.79320 0.46552 0.14677 0.50408 0.78446 0.97226 0.64902 …0.94825 0.75857 7
D14 0.00000 0.33090 0.00000 0.13960 0.00000 0.02756 0.00000 0.00000 0.00000 0.12996 0.53598 0.05831 0.05175 …0.11695 14
D15 0.08826 0.85013 0.00000 0.45815 0.00000 0.45227 0.10317 0.01809 0.07627 0.51394 0.93029 0.32533 0.24143 0.88305 …11
WSEAS TRANSACTIONS on SYSTEMS and CONTROL
DOI: 10.37394/23203.2024.19.43
Blerta (Kristo) Nazarko
E-ISSN: 2224-2856
406
Volume 19, 2024
