Information Technology Reference
In-Depth Information
(i) s −ₒ
. Here
V
( s )
={
0
}
and since
V
( ʘ ) is not empty we clearly have that
V
( s )
Ho V
( ʘ ).
ˉ
−ₒ
ˉ
−ₒ
ʔ for some ʔ . Here
ʘ
( ʘ )
(ii) s
V
( s )
={
1
}
and ʘ
with
V
={
1
}
.
By Lemma 6.32 specialised to full distributions, we have 1
V
( ʘ ). Therefore,
V
( s )
Ho V
( ʘ ).
s ) ʘ . Use Proposition 6.2 to decompose
Now for the general case we suppose ʔ (
ʘ into s ʔ
ʔ ( s )
·
ʘ s such that s
s ʘ s for each s
ʔ
, and recall each
such state s is stable. From above, we have that
V
( s )
Ho V
( ʘ s ) for those s , and so
= ʔ
Ho s ʔ
V
( ʔ )
ʔ ( s )
· V
( s )
ʔ ( s )
· V
( ʘ s )
= V
( ʘ ).
Lemma 6.45
Let ʔ and ʘ be distributions in an ˉ-respecting finitary pLTS given
s ) ʘ, then we have
by
S ,
{
˄ , ˉ
}
,
.Ifʔ (
V
( ʔ )
Ho V
( ʘ ) .
ʔ ;by
distillation of divergence (Theorem 6.11 ) we have full distributions ʔ , ʔ 1 and ʔ 2
and probability p such that s
s )
ʘ , we consider subdistributions ʔ with ʔ
Proof
Since ʔ (
ʔ =
( ʔ 1 p
ʔ 2 ) and ʔ =
ʔ 1 and ʔ 2
p
·
ʘ such that ʔ (
s )
ʘ .By
ʵ . There is, thus, a matching transition ʘ
Proposition 6.2 , we can find distributions ʘ 1 , ʘ 2 such that ʘ =
ʘ 1 p
ʘ 2 , ʔ 1 (
s )
ʘ 1 and ʔ 2 (
s ) ʘ 2 .
ʔ 1 =
ʔ
we have that ʔ 1 is stable. It follows from Lemma 6.44 that
Since
( ʔ 1 )
( ʘ 1 ). Thus we finish off with
V
Ho V
( ʔ )
V
ʔ 1 )
ʔ =
ʔ 1
=
V
( p
·
p
·
( ʔ 1 )
=
p
· V
linearity of
V
( ʘ 1 )
Ho
p · V
above argument based on distillation
ʘ 1 )
=
V
( p
·
linearity of
V
( ʘ )
ʘ =
ʘ 1 p
ʘ 2
Ho
V
Ho
V
( ʘ ) . Lemma 6.32 specialised to full distributions
Since ʔ was arbitrary, we have our result.
Lemma 6.46
Let ʔ and ʘ be distributions in an ˉ-respecting finitary pLTS given
by
S ,
{
˄ , ˉ
}
,
.Ifʔ
S ʘ, then it holds that
V
( ʔ )
Ho V
( ʘ ) .
S ʘ . By Lemma 6.43 , there exists some distribution ʘ match
Proof
Suppose ʔ
ʘ match
s ) ʘ match . By Lemmas 6.45 and 6.32 we obtain
such that ʘ
and ʔ (
( ʘ )
V
( ʔ )
Ho V
V
( ʘ ).
Theorem 6.18
For any finitary processes P and Q,ifP
S Q then P
pmay Q.
Search WWH ::




Custom Search