Escolar Documentos
Profissional Documentos
Cultura Documentos
( )
1
1
1 2
2 1
3
2 3
3 2
3
0,
,
,
0,
A
x a
x a
a x a
a a
x
a x
a x a
a a
x a
<
s s
=
s s
>
( + +
(
(
(
( =
(
(
(
where
( )
1, , 1,
ki
f i A k K = = is the linguistic or numerical value of the optimum criterion
k K e for alternatives ( ) a a A e . If at least one criterion is described by linguistic
expression in the description of the optimum criterion, step 2 is taken.
Step 2. Defining the set of linguistic descriptors. Criterion values are described by a
set of linguistic descriptors , , , ,
1 2
, , , , 0, ,
i
S l l l i H T = . e = . , where T is the overall
number of linguistic descriptors. Linguistic variables are represented by triangular
fuzzy numbers
1 2 3
( , , ) A a a a = .
Step 3. Normalisation of the optimisation criterion. For the value
( )
1, , 1,
ki
f i A k K = = to
be comparable, it is necessary to normalise them. If the optimisation criteria are
given as linguistic values
( )
1, , 1,
ki
f i A k K = = ,
, ,
,
ki
ki ki
f
f f = normalisation is
performed as follows:
- for the benefit criterion ( ) k k K e normalisation is performed as follows:
( )
max
k
k
n
k
i
i
f
f
f
= (1)
where
max
k
f is max value of the fuzzy number
( 1, , )
ki
f k K = . , for
( ) 0
ki
ki
f
f =
- for the cost criterion ( ) k k K e normalisation is performed as follows:
10
( )
min
max
1
ki k
ik
n
k
f f
f
f
= (2)
where
min
k
f is minimum value of the fuzzy number
( 1, , )
ki
f k K = . , for
( ) 0
ki
ki
f
f = .
The normalised value of the criterion ( ) k k K e for ( ) a a A e alternative is described
by fuzzy number:
( )
( )
, ,
,
ik
ik ik
f n
n
f f = (3)
If the optimality criteria are described in numerical values
( )
1, , 1,
ki
f i A k K = = ,
normalisation is performed as follows:
1
1
, 1
k
ki
K
ki
ki ki
k
K
f
f
f f
=
=
= =
(4)
Step 4. Evaluation criteria. , , 1, , , , K k K = . . is a set of optimality criteria, where K
is the overall number of the considered criteria. Every criterion can be disaggregated
into sub-criteria. If
j
k is the overall number of sub-criteria in j
th
criterion the overall
number of criterions can be given as:
1
n
j
j
K k
=
=
(5)
Every criterion has to be divided into sub-criteria. In that case
j
k of the criterion
equals 1. This is important for understanding the aggregation process of judgments
made at two consecutive hierarchical levels, where criteria and sub-criteria are
located. Here criteria and sub-criteria are aggregated by shifting criteria at the sub-
criteria level. After that shift, the whole criteria level does not exist anymore.
Relative importance of the optimality criterion ( ) ( ) , 1, ,
k
k k K W k K e = . is different.
The value representing the importance of the optimality criteria is determined by
forming a matrix
K K
kij
w W
(
=
. Elements of the matrix are linguistic descriptors and
numerical values used to describe the importance of the criterion ( ) k k K e to
criterion ( ) k k K ' ' e . Having established W matrix, normalisation of the weight
coefficients is performed:
11
[
1
1
, 1, 0,1
j
j
K
j k
j
kij k k K
j
k j
w w w
w
w
=
=
= =
(6)
where
j
represents the preference of decision maker to attribute i .
The process of designing the organisational structure is most often in the hands of
more than one expert i.e. decision-maker. In this case, optimality criteria evaluation
of all the group members should first be obtained to pass on to the necessary
synthesis and then to step 5. In other cases, step 6 is taken.
Step 5. Evaluating the criteria in case of group evaluation. In group decision-making,
there is group synthesis with complete and incomplete information.
In case of group synthesis with complete information, provided that all the members
( 1, 2,..., ) e n = of the group G are considered equal in the decision making and that all
the evaluation of the optimality criterion for the given hierarchy have been
performed, there are two ways for prioritising the alternatives relative to the goal.
One is to aggregate all the obtained priority criterion vectors for every decision-
maker using the following equation:
1
( )
n
G
i e i
e
w w e o
=
=
(7)
where ( )
i
w e is the weight value which is the n th member of the group G ,
( 1, 2,..., ) e n = for the alternative
i
A ,
e
o is the weight value (significance) of the n th
member of the group, and
G
i
w is the ultimate priority of the alternative
i
A . Individual
weights of the
e
o group members have first been additively normalised. The
drawback of the presented procedure is that it is not applicable in case of group
synthesis with incomplete information, as there are no composite vectors for certain
members of the group.
Another way is to immediately aggregate all the individual preference assessments
on all hierarchical levels.
1
1 1 1
1 1 1 1
1
j
j
n n n
n n n n
j k
i j ij j ij j ij
i j j j
j
k
k
j
ij K
w
w a a w
w
a
= = = =
=
(
'
( ( (
( ' =
` ( ( (
(
= =
'
)
[ [ [
[
1
1, 0,1
k
j
K
k
w w
=
=
(8)
where
j
represents the preference of the decision-maker to attribute i .
12
N case of group synthesis with incomplete information, microaggregation of the
( ) , i j position at the given matrix is done by geometric mean of the assessments of
those group members who expressed preference
i
E compared to the element
j
E .
The requirement in this case is for at least one decision maker to declare on the
value of
ij
a . Modifying the previous expression:
( )
1
1 1
1
1 1 1
1 j
j
n M
G M
n n n n
j k
G
i j ij j ij j ij kij
i l L j j
j
j
K
G
k
w
w a l a a w
w
= e = =
=
(
( (
(
( =
` ( (
(
(
=
)
=
[ [ [
[
1
1, 0,1
k
j
K
k
w w
=
=
(9)
where l is a set of group members who have evaluated the pair of elements
( )
,
i j
E E ,
and M is the number of such members.
Step 6. Evaluating alternatives. Having determined the values of the weight
coefficients for all the assessed criterions, a matrix
F W
ij
c F
(
=
is formed where the
matrix elements
ij
c are obtained using the following expression:
1
ij
ij k j
ki
ki
i
f
f
c w
=
=
(10)
where
ki
f
is the value of the criterion function for the alternative
( 1, ) i i A =
and
criterion
( ) k k K e
, and
ij
k
w
is the value of the weight coefficient for the criterion
( ) k k K e
.
Additive synthesis has been assumed here and the final alternative performance
weights with respect to overall goal are calculated by the summation of elements in
the rows of the performance matrix
F W
ij
c F
(
=
as:
1
K
i ij j
j
c c w
=
=
(11)
Value of the criterion functions
i
V for every assessed alternative is obtained from the
F matrix using the expression:
1
, ( )
j
K
i j
V c k K
=
= e
(12)
To finally rank the alternatives, the prioritisation of aggregated assessments is
required. Since each
i
V is a triangular fuzzy number, it is necessary to apply the
13
method of ranking triangular fuzzy numbers. There are several methods that can do
this such as the centre of gravity method, the dominance measure method, the o -
cut with interval synthesis method and the total integral value method. The last one
total integral value method (Liou and Wang, 1992), is considered to be a good
choice for performing the task efficiently and, therefore, has been proposed within
this methodology. For the given triangular fuzzy number
1 2 3
( , , ) A a a a = the total
integral value is defined as:
( ) [
1
3 2 1
I (A)= 1 2 , 0,1
T
a a a
( + + e
(13)
In Eq. 13, represents an optimism index which expresses the decision makers
attitude towards risk. A larger value of indicates a higher degree of optimism. In
practical applications, values 0, 0.5 and 1 are used respectively to represent the
pessimistic, moderate and optimistic views of the decision maker.
For given fuzzy numbers A and B it is said that if I (A)<I (B)
T T
, then A B < ; if
I (A) I (B)
T T
= then A B = ; and if I (A) I (B)
T T
> , then A B > .
The final ranking of alternatives means to adopt certain level of optimism of the
decision-maker, then to apply Eq. 13 on fuzzy numbers Eq. 12, and finally to rank
alternatives regarding obtained values for I ( ), 1, ...,
T i
F i N
0, 0
(0.125 ) / 0.125 0 0.125
U
l
x
x x
<
=
`
s s
)
(14)
/ 0.125, 0 0.125
(0.250 ) / 0.125, 0.125 0.25
VL
l
x x
x x
s s
=
`
s s
)
(15.)
s s
=
`
s s
)
(16)
s s
=
`
s s
)
(17)
s s
=
`
s s
)
(18)
s s
=
`
s s
)
(19)
s s
=
`
s s
)
(20)
s s
=
`
s s
)
(21)
s <
=
`
>
)
(22)
To determine the relative importance of the evaluation criteria SWOT, they were
pair-wise compared with respect to the goal by using the fuzzyfied. In Table 1 the
evaluation of linguistic criterions for each of the presented alternatives according to
observed optimality criterions are given.
18
Table 1. Optimality criterion level of influence on the observed alternatives.
Criteria and sub-criteria
Linguistic criteria Benefit-cost criteria
A
1
A
2
A
3
A
4
Min Max
Strengths
1. Capable and competent
personnel
M MH H H -
2. Tactical-operational units swift
dislocation capability
M VH H MH -
3. Strong management team M VH H MH -
Weaknesses
1. Number of hierarchical levels H M M M -
2. Personnel motivation
possibilities
L VH H H -
3. Organisational structure
efficiency
L VH M MH -
4. Coordination VL VH M VH -
5. Sharing experience with
foreign armed forces
VL VH M P -
6. Partial optimisation VH VL L FL -
7. Resources exploitation M VH H VH -
Opportunities
1. Cooperation with foreign
armed forces
L VH MH VH -
2. Liberalisation of personnel
education abroad
M VH MH VH -
3. Modern informational
technologies introduction in the
area of business organisation
L VH M VH -
4. Establishing of logistic support
organisation capable of
satisfying command demands
VL VH M H -
5. Grouping of organisational
units according to NATO
standards
L H M VH -
6. Participation in logistic support
of the NATO forces
VL MH M VH -
Threats
1. Political and economic
instability in Serbia
VH M VH H -
2. Qualified personnel outflow VH MH VH MH -
3. Changes in logistic support
doctrine
VH M H H -
4. Changes in the countrys
defence doctrine
VH M H H -
19
Linguistically expressed preferences among criteria have been used to create a
judgment matrix W as given in step 4.
Following the decision-makers criterion assessment, normalisation of optimality
criteria using Eqs. 1 and 2 is performed. The weighting vector
kij
w
of criteria matrix
W was determined by applying Eq. 6.
Each entry of this vector is the sum of elements in the related row of matrix W ,
divided by the sum of all its elements.
(0.32, 0.30, 0.27) 0.33
(0.40, 0.36, 0.32) 0.29
(0.16, 0.18, 0.21) 0.20
(0.12, 0.15, 0.20) 0.17
s
w
SWOT
o
t
w
w
W
w
w
( ( (
( ( (
( ( (
= = =
( ( (
( ( (
In the next step the judgment matrices for sub-criteria related to respective criteria
were obtained. Related sub-criteria weighting vectors were calculated as defined by
Eqs. 6 (Table 2).
20
Table 2. The priorities of the SWOT factors and groups
SWOT
Local
priority
SWOT factors
Global
priorities
Strengths 0.33
Capable and competent personnel 0.39
Tactical-operational units swift dislocation
capability
0.26
Strong management team 0.35
Weaknesses 0.29
Number of hierarchical levels 0.12
Personnel motivation possibilities 0.13
Organisational structure efficiency 0.14
Coordination 0.17
Sharing experience with foreign armed forces 0.10
Partial optimisation 0.16
Resources exploitation 0.18
Opportunities 0.20
Cooperation with foreign armed forces 0.15
Liberalisation of personnel education abroad 0.18
Modern informational technologies
introduction in the area of business
organisation
0.09
Establishing of logistic support organisation
capable of satisfying command demands
0.20
Grouping of organisational units according to
NATO standards
0.12
Participation in logistic support of the NATO
forces
0.12
Threats 0.17
Political and economic instability in Serbia 0.35
Qualified personnel outflow 0.07
Changes in logistic support doctrine 0.30
Changes in the countrys defence doctrine 0.27
Having determined the value of the weighing coefficients for each of the observed
criterions, matrix
F W
ij
c F
(
=
is formed (Table 3).
21
able 3. Additive synthesis
SWOT factors Alternative 1 Alternative 2 Alternative 3 Alternative 4 W
2
W
1
S
t
r
e
n
g
t
h
s S
1
(0.119,0.113,0.109) (0.119,0.113,0.109) (0.078,0.084,0.088) (0.078,0.084,0.088) 0.39
0,33
S
2
(0.129,0.103,0.092) (0.129,0.103,0.092) (0.000,0.026,0.037) (0.000,0.026,0.037) 0.26
S
3
(0.116,0.108,0.102) (0.116,0.108,0.102) (0.077,0.080,0.083) (0.040,0.053,0.062) 0.35
W
e
a
k
n
e
s
s
e
s
W
1
(0.027,0.028,0.030) (0.027,0.028,0.030) (0.018,0.021,0.024) (0.046,0.042,0.036) 0,12
0,29
W
2
(0.030,0.031,0.032) (0.045,0.042,0.039) (0.030,0.031,0.032) (0.030,0.031,0.032) 0,13
W
3
(0.024,0.027,0.030) (0.036,0.036,0.037) (0.024,0.027,0.030) (0.060,0.053,0.045) 0,14
W
4
(0.016,0.022,0.028) (0.046,0.045,0.046) (0.030,0.033,0.037) (0.076,0.067,0.056) 0,17
W
5
(0.028,0.027,0.027) (0.028,0.027,0.027) (0.018,0.020,0.022) (0.028,0.027,0.027) 0,10
W
6
(0.040,0.040,0.040) (0.040,0.040,0.040) (0.040,0.040,0.040) (0.040,0.040,0.040) 0,16
W
7
(0.025,0.029,0.035) (0.063,0.058,0.052) (0.025,0.029,0.035) (0.063,0.058,0.052) 0,18
O
p
p
o
r
t
u
n
i
t
i
e
s
O
1
(0.017,0.021,0.027) (0.043,0.042,0.040) (0.043,0.042,0.040) (0.043,0.042,0.040) 0,15
0,20
O
2
(0.023,0.028,0.034) (0.058,0.055,0.050) (0.058,0.055,0.050) (0.035,0.037,0.042) 0,18
O
3
(0.012,0.014,0.017) (0.031,0.029,0.026) (0.031,0.029,0.026) (0.012,0.014,0.017) 0,09
O
4
(0.023,0.028,0.036) (0.058,0.056,0.054) (0.058,0.056,0.054) (0.058,0.056,0.054) 0,20
O
5
(0.018,0.021,0.025) (0.046,0.042,0.037) (0.009,0.014,0.019) (0.046,0.042,0.037) 0,12
O
6
(0.014,0.017,0.022) (0.035,0.034,0.032) (0.035,0.034,0.032) (0.035,0.034,0.032) 0,12
T
h
r
e
a
t
s
T
1
(0.050,0.059,0.071) (0.127,0.118,0.106) (0.050,0.059,0.071) (0.127,0.118,0.106) 0,35
0,17
T
2
(0.009,0.011,0.013) (0.022,0.021,0.019) (0.013,0.014,0.016) (0.022,0.021,0.019) 0,07
T
3
(0.076,0.076,0.076) (0.076,0.076,0.076) (0.076,0.076,0.076) (0.076,0.076,0.076) 0,30
T
4
(0.069,0.069,0.069) (0.069,0.069,0.069) (0.069,0.069,0.069) (0.069,0.069,0.069) 0,27
22
where the matrix elements
ij
c
are calculated using Eq. 10.
1 2 3 4
(0.30, 0.32, 0.36) (0.36, 0.32, 0.30) (0.16, 0.19, 0.21) (0.12, 0.16, 0.19)
(0.20, 0.21, 0.23) (0.24, 0.26, 0.23) (0.19, 0.21, 0.23) (0.30, 0.33, 0.35)
(0.07, 0.08, 0.11) (0.29, 0.31, 0.32) (0.26, 0.27, 0.29) (0.26, 0.27, 0.2
A A A A
F =
0.33
0.29
9) 0.20
(0.14, 0.14, 0.17) (0.30, 0.31, 0.32) (0.21, 0.23, 0.25) (0.29, 0.31, 0.34) 0.17
( (
( (
( (
( (
( (
Table 4. Final ranking of alternatives
Decision
alternative
Index of optimism
Final rank
=0.0
(pessimistic)
=0.5
(moderate)
=1.0
(optimistic)
Alternative 1 0.205 0.212 0.220 4
Alternative 2 0.285 0.295 0.305 1
Alternative 3 0.210 0.220 0.230 3
Alternative 4 0.245 0.255 0.265 2
The assessment of alternatives has been performed using relations 10 and 12. The
final alternative performance weights, with respect to the overall goal, have been
calculated by Eq. 32 as:
1
2
3
4
(0.20, 0.21, 0.23)
(0.28, 0.29, 0.32)
(0.20, 0.22, 0.24)
(0.22, 0.26, 0.27)
V
V
V
V
V
( (
( (
( (
= =
( (
( (
For the typical values of that express the decision-makers attitude toward risk, the
final ranking of alternatives is obtained by applying Eq. 13. The normalised values
presented in Table 4 show that Alternative 2 is the best. It is followed by Alternative
4, Alternative 3 and Alternative 1 respectively, regardless of the decision-makers
level of optimism.
By using the centre of gravity method to defuzzify the V values given above, the
final weights of alternatives obtained after normalization were: 0.213 (Alternative 1),
0.297 (Alternative 2), 0.220 (Alternative 3) and 0.253 (Alternative 4). Obviously, the
final ranking is equal to the previous one.
5. DISCUSSION AND CONCLUSION
Organisation is not a sum of mechanical parts, rather an "organic whole" with a
purpose and mission. In the process of designing the organisational structure it is
23
necessary, having defining the objectives and design criteria, to analyse the state of
the organisation.
In addition to organisations operating in an uncertain environment, there is a degree
of uncertainty and imprecision of criteria used in the process of organisational
design. Fuzzy multi-criteria approach developed in this paper allows the
quantification of these criteria and selection of the best alternative out of proposed
organisational models. The presented model enables the evaluation of the proposed
options of organisational structure, regardless of the number of optimality criteria and
sub-criteria. The model allows for the choice of best alternative from a set described
using K optimality criteria and sub-criteria. The criteria described can be of benefit
or cost type. The criteria relevant to the design of organisations as well as their
influence on the choice of alternatives have their values displayed as numerical
values or fuzzy linguistic descriptors. Since the process of organisational design
often involved a number of experts, the model allows for a possibility of optimality
criteria values synthesis in case of group decision-making. Decision-making in a
group differs from individual decision-making on methodological and mathematical
levels. Group syntheses with complete and incomplete information are discussed in
the model. In addition, the model enables comparison of the criterion functions
output values using two methods dephasification of the centre of gravity and the total
integral value method.
Application of the given model is shown on the example of designing the
organisational structure of the governing bodies of transport support. The complex
environment, in which these governing bodies act, does not tolerate organisational
improvisation, rather requires a planned and methodological organisational project
and its constant modification and adaptation. Selection of the appropriate
organisational structure is one of the most significant decisions, as the capabilities of
the governing system will be significantly slowed down if an organisational structure
is inadequate for the circumstances in which the organisation is.
Although the model application was shown on the example of designing governing
bodies within the armed forces, it possesses great flexibility and can be adapted to
any particular problem. Very easily, with minor modifications, it can be applied for the
selection of organisational structure of any business system.
6. ACKNOWLEDGEMENT
24
This paper is a part of the project funded by the Ministry of defence of the Republic
of Serbia and University of defence of the Republic of Serbia, Project number:
T3/11013.
REFERENCES
Bender MJ, Simonovic SP (2000). A fuzzy compromise approach to water resources
systems planning under uncertainty. Fuzzy Set. Syst.,115: 35-44.
Bozanic D, Pamucar D (2010). Evaluating locations for river crossing using fuzzy
logic. Milit. Techn. Cour., 1: 129-145.
Brans JP, Vincke P, Mareschal B (1986). How to select and how to rank projects by
the Promethee method. Eur. J. Oper. Res., 24: 228-238.
Deng H (1999). Multicriteria analysis with fuzzy pairwisecomparison. Int. J. Approx.
Reason., 21: 215-231.
Hwang CL, Yoon KS (1981). Multiple attribute decision making: methods and
applications. Springer, Berlin.
Kujacic M, Bojovic N (2003). Organizational design of post corporation structure
using fuzzy multicriteria decision making. Comput. Mathematical Organ. Theory, 9:
5-18.
Liou TS, Wang MJJ (1992). Ranking fuzzy numbers with integral value. Fuzzy Set.
Syst., 50: 247-256.
Pamucar D (2009). Design of the organisational structure using fuzzy logic
approach. Master paper. Serbia: Faculty of Transport and Traffic Engineering,
Belgrade.
Pamucar D (2010). Using fuzzy logic and neural networks during decision making
proces in transport. Milit. Techn. Cour., 3: 125-143.
Pamucar D, Bozanic D, Dorovic B, Milic A (2011a). Modelling of the fuzzy logical
system for offering support in making decisions within the engineering units of the
Serbian army. Int. J. Phys. Sci, 3: 592-609.
Roy B (1968). Ranking and selection in the presence of multiple viewpoints. Comput.
Oper. Res., 8: 57-75.
Saaty TL (1980). The analytic hierarchy process. McGraw-Hill, New York.
Srdevic B, Medeiros Y, Srdevic Z, Schaer M (2002). Evaluating management
strategies in Paraguacu river basin by analytic hierarchy process. First Biennial
25
Meeting of the International Environmental Modeling and Software Society, 1: 42-
47.
Triantaphyllou E, Lin CT (1996). Development and evaluation of five multiattribute
decision making methods. Int. J. Approx. Reason., 14: 281-310.
Zadeh LA (1965). Fuzzy sets. Inf. Control. 8: 338-353.
Zeleny M (1982). Multiple citeria decision making. McGraw-Hill Book Company, New
York.