Graphics Programs Reference
In-Depth Information
t
2
p
in
t
1
p
1
p
2
T
1−2
p
3
T
3
p
out
/
2
2
p
4
p
6
T
4
t
3
t
4
t
5
p
5
p
7
t
6
p
control
Figure 7.9: Enabling memory conflict
firing of transition t
1
and is returned empty either on completion - firing
of T
3
- or at the end of the clearing action we are discussing). As soon as
place p
in
becomes empty, immediate transitions t
3
, t
4
, and t
5
fire as many
times as is needed to remove all the tokens that may be found in these
three places. At the end of this first step, places p
4
and p
5
are marked
with one token each. The return to the initial state is finally performed
by transition t
6
that puts one token in place p
control
, making sure that this
action takes place only after places p
1
, p
2
, and p
3
are empty using inhibitor
arcs. The part of the GSPN system that we have just described appears
cumbersome, but has the advantage of confining within the subnet itself the
entire “mechanism” that is needed to lose memory of the work performed
before the interruption. To make sure that these actions take place in the
proper sequence, different priority levels may also be assigned to transitions
t
3
, t
4
, t
5
, and t
6
. In particular, transitions t
3
, t
4
, and t
5
may be defined of
priority level 3 and transition t
6
of priority level 2.
tures of being of the one-input/one-output type and of having the specifica-
tions of the race memory policies completely “self-contained”. This makes
possible the definition of a reduction/expansion rule that allows the auto-
matic interchanging of simple timed transitions with more complex subnets
that represent their internal expansions and vice versa. Indeed, the two
Search WWH ::
Custom Search