A resource place in a net system is modeled with a
singleton place whose token capacity represents the
ability of processing raw parts concurrently.
Example. From the definition of a G-system, there are
four minimal P-semiflows that can be computed in
Figure 1(b) shown as below, which is associated with
resource places p
15
p
17
, respectively
I
p
15
= 2p
1
+ p
12
+ p
15
, H(p
15
)=2p
1
+ p
12
I
0
p
15
= 2p
5
+ p
12
+ p
15
, H(p
15
)
0
= 2p
5
+ p
12
,
M
0
(p
15
)=2
I
p
16
= p
2
+ p
6
+ p
11
+ p
16
, H(p
16
)=p
2
+ p
6
+ p
11
, M
0
(p
16
)=2
I
p
17
= p
3
+ p
4
+ p
7
+ p
10
+ p
17
,
H(p
17
)=p
3
+ p
4
+ p
7
+ p
10
, M
0
(p
17
)=3 }
Definition 4. Let p
j, k
be an operation place of J
j
in
GS =(N
S
, M
S
0
, M
S
F
). The resource allocation require-
ment of p
j, k
is defined as a jP
R
j-dimensional vector
a
p
j, k
=(a
p
j, k
½r
1
, ..., a
p
j, k
½r
jP
R
j
)
T
. a
p
j, k
½r
i
= I
r
i
(p
j, k
) indi-
cates how many units of r
i
are needed in the operation
p
j, k
.
Example. Take the net shown in Figure 1(b) as an
example. For operation place p
4
, p
4
is only included in
P-semiflow I
p
17
, that is to say, resource place p
17
can
support p
4
, then a
p
4
=(0, 0, 1)
T
can be derived.
Similarly, the net has two job types J
1
= fp
1
p
7
g and
J
2
= fp
10
p
12
g. From above minimal P-semiflows,
8p
i
2 P
A
, its related resources requirements a
p
i
in J
1
and J
2
can be computed as shown in Table 1.
By considering a G-system under different resource
requirements, monitors should be added to the net sys-
tem in order to prevent the controlled system entering
into the forbidden markings, the policy under resource
partial orders is presented as follows.
Definition 5. Let o
i
[O(r
i
), O : P
R
! N
jP
R
j
be any partial
order imposed on the resource places set P
R
. Given an
operation place p
j, k
2 P
A
j
, where j 2 N
n
, k 2 N
jP
A
j
j
, let
r
max
p
j, k
=
maxfo
i
ja
p
j, k
½i.0, i 2 N
jP
R
j
g if 9r 2 P
R
, p
j, k
2k H(r) k
0 otherwise
ð1Þ
and
r
min
p
j, k
=
minfo
i
ja
p
j, k
½i.0, i 2 N
jP
R
j
g if 9r 2 P
R
, p
j, k
2k H(r) k
0 otherwise
ð2Þ
Let L
p
j, k
= fqjq 2 p
j, k
\ P
A
j
^ r
max
q
= min
v2p
j, k
\P
A
j
r
max
v
g, by convention, L
p
j, k
= [ if p
j, k
\ P
0
6¼ [.
Then,
1. The neighborhood set N
p
j, k
of p
j, k
2 P
A
j
is
defined recursively by the following equation
N
p
j, k
= fp
j, k
g[fqjq 2[
v2L
p
j, k
N
v
^ r
min
p
j, k
r
max
q
g
2. 8p
j, k
2 P
A
j
, the adjusted resource requirement
^
a
p
j, k
is defined as an non-negative jP
R
j-dimen-
sional vector due to O(r
i
), which is computed as
below
^
a
p
j, k
½i =
maxfa
q
½ijq 2 N
p
j, k
g
0
if o
i
r
min
p
j, k
, i 2 N
jP
R
j
otherwise
ð3Þ
Definition 5 formalizes the liveness control specifica-
tion of resources requirements, which can be expressed
by a set of generalized mutual exclusion constraints
(GMECs). The neighborhood set N
p
j, k
of an operation
place p
j, k
is defined at Definition 5(1), which indicates
that the maximum resource order of the resources used
by an operation q in N
p
j, k
is no less than the minimum
one used by p
j, k
. Under a certain resources partial
order, Definition 5(2) presents a formalized method on
computing the adjusted resource requirements of opera-
tion places, which indicates that the appropriate adjust-
ment of system resources should be carried out such
that each operation in its neighborhood set are properly
allocated.
Table 1. The resources requirements in P
A
for the net system in Figure 1(b).
P
R
a
p
1
a
p
2
a
p
3
a
p
4
a
p
5
a
p
6
a
p
7
a
p
10
a
p
11
a
p
12
p
15
20002000 0 1
p
16
01000100 1 0
p
17
00110011 0 0
4 Advances in Mechanical Engineering