Information Technology Reference
In-Depth Information
Therefore,
ʔ ( A )
inf
n 0 ʘ (
n ( A ))
=
ʘ ((
n 0 n )( A ))
=
ʘ (
ˉ ( A )) .
Using Theorem 3.5 (1) again, we obtain ( 3.12 ).
By symmetry we also have that if t
a
−ₒ
a
−ₒ
ʘ then there is some ʔ with s
ʔ
ˉ ) ʘ .
and ʔ (
Let
be the largest probabilistic simulation, called probabilistic similarity, and
≺∩≺ 1 , called simulation equivalence .
In general, simulation equivalence is coarser than bisimilarity. However, for reactive
pLTSs, the two relations do coincide. Recall that in a reactive pLTS, for each state s
and action a there is at most one distribution ʔ with s
be the kernel of probabilistic similarity, i.e.
a
−ₒ
ʔ . To prove that result,
we need a technical lemma.
Lemma 3.3
Let
R
be a preorder on a set S and ʔ , ʘ
D
( S ) .Ifʔ
R
ʘ and
ʔ then ʔ ( C )
ʘ
R
=
ʘ ( C ) for all equivalence classes C with respect to the kernel
R R 1
of
R
.
R R 1 . For any s
Proof
Let us write
for
S , let [ s ] be the equivalence class
{
|
R
R
}
that contains s . Let A s be the set
t
S
s
t
t
s
. It is easy to see that
R
( s )
={
t
S
|
s
R
t
}
={
t
S
|
s
R
t
t
R
s
}{
t
S
|
s
R
t
t
R
s
}
=
[ s ]
A s
where
stands for a disjoint union. Therefore, we have
ʔ (
R
( s ))
= ʔ ([ s ] )
+ ʔ ( A s )
and
ʘ (
R
( s ))
= ʘ ([ s ] )
+ ʘ ( A s ) . (3.13)
We now check that both
R
( s ) and A s are
R
-closed sets, that is
R
(
R
( s ))
R
( s )
and
R
( A s )
A s . Suppose u
R
(
R
( s )). Then there exists some t
R
( s ) such that
t
R
u , which means that s
R
t and t
R
u . As a preorder
R
is a transitive relation. So we
have s
R
u that implies u
R
( s ). Therefore we can conclude that
R
(
R
( s ))
R
( s ).
Suppose u
R
( A s ). Then there exists some t
A s such that t
R
u , which means
that s
R
t , t
R
s and t
R
u . As a preorder
R
is a transitive relation. So we have
s
R
u . Note that we also have u
R
s . Otherwise we would have u
R
s , which
means, together with t
R
u and the transitivity of
R
, that t
R
s , a contradiction to the
hypothesis t
R
s . It then follows that u
A s and then we conclude that
R
( A s )
A s .
ʘ and ʘ
ʔ ,
We have verified that
R
( s ) and A s are
R
-closed sets. Since ʔ
R
R
R
R
R
we apply Theorem 3.5 (2) and obtain that ʔ (
( s ))
ʘ (
( s )) and ʘ (
( s ))
R
ʔ (
( s )), that is
ʔ (
R
( s ))
=
ʘ (
R
( s )) .
(3.14)
Similarly, using the fact that A s is
R
-closed we obtain that
ʔ ( A s )
= ʘ ( A s )
(3.15)
Search WWH ::




Custom Search