Download - Academy Publisher
Download - Academy Publisher
Download - Academy Publisher
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
symbolic aggregation operation. β ∈[ 0, g]<br />
and g+1 is<br />
the cardinality of S. Let i = round(β ) and α = β − i<br />
be two values, such that, i ∈ [ 0, g]<br />
and<br />
α ∈[−0.5,0.5] then α is called a Symbolic<br />
Translation.<br />
Let S={s0, s1, …,sg} be a linguistic term set and<br />
β ∈[ 0, g]<br />
be a value representing the result of a<br />
symbolic aggregation operation, then the 2-tuple that<br />
expresses the equivalent information to β is obtained<br />
with the following function:<br />
∇ :[0, g]<br />
→ S × [ −0.5,0.5]<br />
⎧ si<br />
, i = round(<br />
β)<br />
(1)<br />
∇(<br />
β)<br />
= ( si<br />
, α),<br />
with⎨<br />
⎩α<br />
= β − i,<br />
α ∈[<br />
−0.5,0.5]<br />
Where round(.) is the usual round operation, si had the<br />
closest index label to β .<br />
Proposition 1[14] Let S={s0, s1, …,sg}be a<br />
linguistic term set and ( s i<br />
, α)<br />
be a 2-tuple. There is<br />
−1<br />
always a ∇ function, such that, from a 2-tuple it<br />
returns its equivalent numerical value β ∈[ 0, g]<br />
, which<br />
is:<br />
−1<br />
∇ : S × [ −0.5,0.5]<br />
→ [0, g]<br />
(2)<br />
−1<br />
∇ ( si , α ) = i + α = β<br />
⑴ A linguistic 2-tuple negation operator [14]<br />
−1<br />
Neg((<br />
si<br />
, α))<br />
= ∇(<br />
g − ( ∇ ( si<br />
, α)))<br />
(3)<br />
⑵ Linguistic 2-tuple aggregation operators<br />
Let ( s1,<br />
α1),(<br />
s1,<br />
α<br />
2<br />
), L ,( s n<br />
, α<br />
n<br />
) be a set with n<br />
linguistic 2-tuples and ω = ( ω1,<br />
ω2<br />
, L,<br />
ωn<br />
) be the<br />
n<br />
related weighted vector with ∑ω<br />
i<br />
= 1 , then the<br />
i=<br />
1<br />
weighted average operator of linguistic 2-tuples<br />
ξ ω<br />
= ∇(<br />
(( s , α ),( s<br />
1<br />
n<br />
∑<br />
i=<br />
1<br />
1<br />
∇<br />
−1<br />
2<br />
, α ), L,(<br />
s<br />
2<br />
( s , α ) ω )<br />
i<br />
i<br />
i<br />
n<br />
ω<br />
ξ is<br />
, α )) = (ˆ, s ˆ) α<br />
Let P ( pij, α<br />
ij<br />
)<br />
n × n<br />
comparison matrix and the element ( p , α )<br />
n<br />
(4)<br />
= be a linguistic 2-tuple<br />
ij<br />
ij<br />
represent<br />
the result of comparing two solutions. If the following<br />
propositions are right.<br />
(1) p ∈S; α ∈[ −0.5,0.5]<br />
ij ij<br />
−1<br />
(2) ∇ ( pii, αii<br />
) = g/ 2<br />
−1 −1<br />
ij<br />
αij ji<br />
α<br />
ji<br />
(3) ∇ ( p , ) +∇ ( p , ) = g<br />
then P is called a linguistic 2-tuple judgment matrix.<br />
(5)<br />
Let P= ( pij , α<br />
ij<br />
)<br />
n × n<br />
be a linguistic 2-tuple judgment<br />
matrix, if ∀ i, j,<br />
k ∈ I , elements in P has the properties<br />
of the formula (6), then P is called a linguistic 2-tuple<br />
judgment matrix with additive consistency.<br />
−1 −1 −1<br />
∇ ( p , α ) +∇ ( p , α ) =∇ ( p , α ) + g/2, ∀i, j,<br />
k∈ I (6)<br />
ij ij jk jk ik ik<br />
Ⅲ. CALCULATE THE ELEMENT CONSISTENCY LEVEL<br />
If the judgment matrix given by an expert is additive<br />
consistent, the comparison of each pair of alternative is<br />
identical with the indirect value based on additive<br />
consistency. In real decision making environment,<br />
however, one element in the given judgment matrix may<br />
have high similarity to its indirect value and another<br />
element may have low similarity to its indirect value.<br />
Therefore, it is unreasonable to give the expert a fixed<br />
weight when the judgment matrices are aggregated into<br />
group decision judgment matrix. To measure the<br />
similarity between an element and its indirect valued, the<br />
concept of element consistency level was introduced .<br />
If a linguistic 2-tuple representation judgment matrix P<br />
P = ( , α ) ) is additive consistent, there exists<br />
p ij<br />
( [<br />
ij<br />
] n × n<br />
∇<br />
−1<br />
( p<br />
−1<br />
−1<br />
, α ) = ∇ ( p , α ) − ∇ ( p , α ) g / 2 .<br />
ij ij<br />
kj kj<br />
ki ki<br />
+<br />
The property can be used to compute the indirect value of<br />
an element in the judgment matrix.<br />
Based on the two properties of an additive consistent<br />
linguistic 2-tuple judgment matrix<br />
−1<br />
−1<br />
−1<br />
∇ ( pij , α<br />
ij<br />
) = ∇ ( pkj<br />
, α<br />
kj<br />
) − ∇ ( pki<br />
, α<br />
ki<br />
) + g / 2<br />
−1<br />
−1<br />
and ∇ ( pij , α<br />
ij<br />
) + ∇ ( p<br />
ji<br />
, α<br />
ji<br />
) = g , the following<br />
formulas can be reasoned out.<br />
−1<br />
∇ ( p , α ) =∇<br />
ij<br />
ij<br />
ij<br />
ij<br />
ij<br />
ij<br />
−1<br />
−1<br />
∇ ( p , α ) =∇<br />
−1<br />
−1<br />
∇ ( p , α ) =∇<br />
−1<br />
( p<br />
( p<br />
kj<br />
( p<br />
ik<br />
ik<br />
, α ) +∇<br />
ik<br />
, α ) −∇<br />
kj<br />
, α ) −∇<br />
ik<br />
−1<br />
−1<br />
−1<br />
( p<br />
( p<br />
ki<br />
( p<br />
kj<br />
, α ) −g/2<br />
, α ) + g/2<br />
jk<br />
kj<br />
ki<br />
, α ) + g/2<br />
Therefore, the indirect valued of an element in a<br />
linguistic 2-tuple judgment with additive consistency can<br />
be calculated through the neighbor elements. There are<br />
different elements in the judgment matrix can be used to<br />
compute an element’s indirect value, thus, to assessment<br />
the indirect values comprehensively, the RMM( Row<br />
Mean Method) is used to calculate the indirect value of<br />
an element. The following formulas is induced from the<br />
above formula<br />
cp<br />
cp<br />
p1<br />
ij<br />
p2<br />
ij<br />
cp<br />
=<br />
n<br />
∑<br />
k = 1<br />
k ≠i,<br />
j<br />
=<br />
p3<br />
ij<br />
( ∇<br />
n<br />
∑<br />
k=<br />
1<br />
k≠i,<br />
j<br />
=<br />
−1<br />
( ∇<br />
n<br />
∑<br />
k=<br />
1<br />
k≠i,<br />
j<br />
( p<br />
−1<br />
( ∇<br />
p<br />
ik<br />
( p<br />
−1<br />
p<br />
, α<br />
p<br />
kj<br />
( p<br />
ik<br />
p<br />
, α<br />
p<br />
ik<br />
) + ∇<br />
−1<br />
n − 2<br />
kj<br />
p<br />
, α<br />
( p<br />
) −∇<br />
−1<br />
n − 2<br />
ik<br />
p<br />
kj<br />
( p<br />
) −∇<br />
−1<br />
n − 2<br />
p<br />
, α<br />
p<br />
ki<br />
( p<br />
kj<br />
jk<br />
) − g / 2)<br />
p<br />
, α<br />
p<br />
jk<br />
ki<br />
) + g /2)<br />
p<br />
, α<br />
ik<br />
) + g /2)<br />
(7)<br />
(8)<br />
(9)<br />
(10)<br />
70