Information Technology Reference
In-Depth Information
·
Proof
The operations min, and max are obviously idempotent. Let us show that if
·=
μ
·
˃
(μ, ˃)
is idempotent, it must be
min. By Theorem
2.2.6
, it is always
min
,
and the idempotency of
·
implies
min
(μ, ˃)
·
min
(μ, ˃)
=
min
(μ, ˃).
But min
(μ, ˃)
μ,
min
(μ, ˃)
˃
, imply min
(μ, ˃)
·
min
(μ, ˃)
μ
·
˃
, that is
min
(μ, ˃)
μ
·
˃,
and, by Theorem
2.2.6
,min
(μ, ˃)
=
μ
·
˃
. A similar proof applies to
+
and
max
.
Theorem 2.2.8
(Absortion laws)
•
μ
·
(μ
+
˃)
=
μ
X
holds for all
μ, ˃
∈[
0
,
1
]
⃔·=
min
X
•
μ
+
(μ
·
˃)
=
μ
holds for all
μ, ˃
∈[
0
,
1
]
⃔+=
max
·=
(μ, μ
+
˃)
=
μ
μ
μ
+
˃
Proof
If
min, the formula min
does hold, since
.
μ
·
(μ
+
˃)
=
μ
˃
=
μ
0
follows
μ
·
(μ
+
μ
0
)
=
μ
·
μ
=
μ
Provided it is always
, taking
,
that holds if and only if
·=
min. If
+=
max, the formula max
(μ, μ
·
˃)
=
μ
does
hold since
μ
·
˃
μ
. Provided it is always
μ
+
(μ
·
˃)
=
μ
, taking
˃
=
μ
1
follows
μ
+
(μ
·
μ
1
)
=
μ
+
μ
=
μ
, that holds if and only if
+=
max.
(Duality, or De Morgan laws)
Provided the
complement
Theorem 2.2.9
is invo-
((μ
)
=
μ
X
X
lutive
=
μ
, for all
μ
∈[
0
,
1
]
)
, the algebra of fuzzy sets
(
[
0
,
1
]
,
,
)
min
,
max
is a dual algebra.
Proof
If
is involutive, from
μ
+
˃
=
(μ
·
˃)
it follows
μ
·
˃
=
(μ
+
˃)
since
μ
+
˃
=
μ
+
˃
=
(μ
·
˃
)
, hence
(μ
+
˃)
=
μ
·
˃
. The converse is proven in
the same way. And the two De Morgan laws
(μ
·
˃)
=
μ
+
˃
, μ
+
˃)
=
μ
·
˃
(μ
, ˃
))
, for all
result equivalent. Then, it is enough to prove max
(μ, ˃)
=
(
min
X
. Since
μ, ˃
in
[
0
,
1
]
(μ
, ˃
)
μ
,
(μ
, ˃
)
˃
,
min
min
it follows
μ
=
μ
(
(μ
, ˃
))
,
˃
=
˃
(
(μ
, ˃
))
,
min
min
and
(μ
, ˃
))
max
(μ, ˃)
(
min
(2.1)
On the other hand,
(μ, ˃))
μ
μ
(μ, ˃)
⃒
(
max
max
Search WWH ::
Custom Search