Information Technology Reference
In-Depth Information
Definition 1.
A binary relation
R
⊆S×S
is a bisimulation if for all
(
α, β
)
∈
R
the following hold:
1.
α
=
if and only if
β
=
;
2. whenever
α
a
−→
a
−→
α
,then
β
β
for some
β
with
(
α
,β
)
∈
R
;
a
−→
a
−→
β
,then
α
α
for some
α
with
(
α
,β
)
3. whenever
β
∈
R
.
Two states
α
and
β
are said to be bisimulation equivalent, written
α
∼
β
,if
there is a bisimulation
R
such that
(
α, β
)
∈
R
.
It is well know that
is an equivalence relation between processes. Moreover it
is a congruence with respect to concatenation of sequences of processes.
Note that in general clause 1. of Definition 1 is necessary. Since here we use
the general summation notation
Σ
to write expressions, and when the index set
is empty,
Σ
i∈∅
E
i
is the inactive process which should not be equated with
.
Otherwise the congruence property of
∼
∼
would fail. With clause 1. in Definition
1 this cannot happen.
An important notion in the study of bisimulation of context-free processes is
that of self-bisimulation proposed by Caucal [Cau90].
,wedenoteby
R
→
the least
precongruence wrt sequential composition containing
R
,denoteby
R
↔
the
symmetric closure of
R
→
,denoteby
R
↔
∗
the reflexive and transitive closure
of
R
↔
.
Definition 2.
For any binary relation
R
on
S
Definition 3.
A binary relation
B
⊆S×S
is a self-bisimulation if for all
(
α, β
)
∈
B
,
α
=
just in case
β
=
and moreover the following hold:
a
−→
a
−→
B
↔
∗
;
α
,then
β
β
for some
β
with
(
α
,β
)
1. whenever
α
∈
a
−→
a
−→
B
↔
∗
.
β
,then
α
α
for some
α
with
(
α
,β
)
2. whenever
β
∈
Note that for a finite relation, checking if it is a self-bisimulation is in general only
semi-decidable since one needs to check membership of its congruence closure.
Proposition 1.
[Cau90]
If
B
is a self-bisimulation then
B
↔
∗
⊆∼
.
B
↔∗
.Thusif(
α, β
) is contained in a self-bisimulation than
Obviously
B
⊆
α
∼
β
.
For the process system
, an important finiteness property which
forms the base of our discussion in the paper is that, although the number
of reachable states from an arbitrary state
α
S
,Act,
−→
can be infinite, the processes
(closed expressions) that can occur in those reachable states must be from a finite
set. Another important finiteness property of
∈S
S
,Act,
−→
is finite branching,
which says that for any
α
∈S
,theset
a
−→
B
(
α
)=
{
β
|
a
∈
Act, α
β
}
Search WWH ::
Custom Search