Information Technology Reference
In-Depth Information
=
(
) =
(
(
),
(
))
(
)
Since A
B
A
B
A ,itis N
A
min
N
A
N
B
N
B
.
(
(
),
(
))
Hence, all these mappings are actually fuzzy measures. From min
N
A
N
B
A C
ˀ(
A
) + ˀ(
B
)
,itfollows N
(
A
B
) ˀ(
A
) + ˀ(
B
)
, Then 0
=
N
( ) =
N
(
A
) =
A C
min
(
N
(
A
),
N
(
))
, that implies:
A C
A C
Either N
(
A
) =
0
,
or N
(
) =
0
,
and N
(
A
) +
N
(
)
1
,
or
A C
N
(
)
1
N
(
A
).
Obviously, if A 1 ,
A 2 ,...,
A n
∈ F :
N
(
A 1
A 2 ...
A n ) =
min
(
N
(
A 1 ),
N
(
A 2 ),...,
N
(
A n ))
. Then if X
={
x 1 ,...,
x n }
, to define a necessity measure it is
needed to take N
(
x 1 ),
N
(
x 2 ),...,
N
(
x 3 )
such that: 0
=
min
(
N
(
x 1 ),
N
(
x 2 ),...,
N
(
x n ))
, an equality that forces some of the values N
(
x i )
to be 0.
Theorem 7.5.4
Given a possibility measure
ˀ : F ₒ[
0
,
1
]
, the function N ˀ (
A
) =
A C
1
ˀ(
)
, for all A
∈ F
, is a necessity measure.
Proof N ˀ ( ) =
1
ˀ(
X
) =
0. N ˀ (
X
) =
1
ˀ( ) =
1. N ˀ (
A
B
) =
A C
B C
A C
B C
A C
B C
1
ˀ(
) =
1
max
(ˀ(
), ˀ(
)) =
min
(
1
ˀ(
),
1
ˀ(
)) =
(
N ˀ (
),
N ˀ (
))
min
A
B
.
Theorem 7.5.5
Given a necessity measure N
: F ₒ[
0
,
1
]
, the function
ˀ N (
A
) =
A C
1
N
(
)
, for all A
∈ F
is a possibility measure.
Proof
ˀ N ( ) =
1
N
(
X
) =
0.
ˀ N (
X
) =
1
N
( ) =
1.
ˀ N (
A
B
) =
1
A C
B C
A C
B C
N
(
) =
1
min
(
N
(
),
N
(
)) =
max
N (
A
), ˀ N (
B
))
.
are called dual-pairs of possibility/necessity mea-
sures. Notice that with them it can be read:
The pairs
(ˀ,
N
ˀ )
and
(
N
, ˀ N )
N ecessit y o f A
=
not the possibility of not A
Possibility of A
=
not the necessit y o f not A
.
Remark 7.5.6 If
ˀ = ˀ μ
, the corresponding N
is given by
ˀ μ
N ˀ μ (
A
) =
1
ˀ μ (
A
) =
1
Sup
x
min
(μ(
x
), μ A (
x
)) =
Inf
x
max
(
1
μ(
x
), μ A C
(
x
)),
X
X
that, in the finite case X
={
x 1 ,...,
x n }
,is
N ˀ μ (
) =
(
μ(
x i ), μ A C
(
x i )).
A
min
1
i n max
1
Theorem 7.5.7
For all dual pair
(ˀ,
N
)
is:
1.
If N
(
A
)>
0 , then
ˀ(
A
) =
1
2.
If
ˀ(
A
)<
1 , then N
(
A
) =
0
Search WWH ::




Custom Search