Download - Academy Publisher
Download - Academy Publisher
Download - Academy Publisher
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
(2) T ' = T ∪{<br />
~<br />
t } − Tt<br />
;<br />
~ • ~<br />
•<br />
(3) F ' = F ∪{(<br />
p,<br />
t ) | p ∈ ti}<br />
∪{(<br />
t , p)<br />
| p ∈to<br />
}<br />
•<br />
o<br />
} − Ft<br />
( , M<br />
0<br />
•<br />
−{( p,<br />
ti)<br />
| p∈<br />
ti}<br />
−{(<br />
to,<br />
p)<br />
| p ∈t<br />
.<br />
Definition 4.3 ( N ', M<br />
0'<br />
) obtained from N ) by<br />
t-subnet reduction operation comprises net N'<br />
and<br />
marking M<br />
0'<br />
, where M 0' = M<br />
( P \ Pt<br />
) 0<br />
(where M<br />
( P\<br />
Pt<br />
)<br />
is obtained from M by deleted the vector corresponding<br />
to P<br />
t<br />
).<br />
Definition 4.4 A net ( N<br />
t<br />
, M t 0)<br />
is said to be a t-closed<br />
net if we add a transition t<br />
t<br />
and arcs<br />
•<br />
•<br />
{( p , t t<br />
) | p ∈t<br />
o<br />
} ∪{( tt<br />
, p)<br />
| p ∈ ti}<br />
to t-subnet<br />
( N<br />
t<br />
, M t 0)<br />
, and the marking of ( N<br />
t<br />
, M t 0)<br />
is preserved.<br />
Note that in this section, let<br />
( N ', M<br />
0<br />
') : the original net; N<br />
t<br />
= ( Pt<br />
, Tt<br />
; Ft<br />
, Wt<br />
) : the<br />
p-subnet ; ( N<br />
t<br />
, M t 0)<br />
: the p-subnet system; ( N<br />
t<br />
, M t 0)<br />
:<br />
the closed t-subnet system; ( N , M<br />
0<br />
) : the reduced net.<br />
Theorem 4.1 Suppose that ( N , M<br />
0)<br />
is obtained from<br />
(<br />
0<br />
N ', M ')<br />
by t-subnet reduction operation. Then<br />
( N ', M<br />
0'<br />
) is bounded iff ( N,<br />
M<br />
0)<br />
and ( N<br />
t<br />
, M t 0)<br />
are<br />
bounded.<br />
Proof. (1) Since N , M ) is bounded, then ∀ p ∈ P ,<br />
(<br />
0<br />
∃k > 0 1<br />
such that ∀ M ∈ R ( M<br />
0)<br />
, M ( p)<br />
≤ k1<br />
.<br />
Since ( N<br />
t<br />
, M t 0)<br />
is bounded, then ∀ p ∈ Pt<br />
, ∃k > 2<br />
0<br />
such that ∀ M<br />
t<br />
∈ R( M t 0)<br />
, M t<br />
( p)<br />
≤ k2<br />
.<br />
Let k = k 1<br />
+ k2<br />
, ∀ p ∈ P'<br />
, ∀ M ' ∈ R(<br />
M<br />
0'<br />
) such that<br />
M '(<br />
p)<br />
≤ k . So, ( N ', M<br />
0'<br />
) is bounded.<br />
(2) Suppose that N , M ) is unbounded, then<br />
(<br />
0<br />
∃ p ∈ P , ∀k > 0 , M ∈ R M )<br />
∃ such that<br />
(<br />
0<br />
' ( M<br />
0<br />
M ( p)<br />
> k . That is ∀k > 0 , M ∈ R ')<br />
that M '(<br />
p)<br />
> k<br />
N ', M ') is bounded.<br />
(<br />
0<br />
∃ such<br />
. This contradicts with the fact that<br />
Theorem 4.2 Suppose that N , M ) is obtained from<br />
(<br />
0<br />
(<br />
0<br />
N ', M ')<br />
by t-subnet reduction operation. If<br />
•<br />
( N ', M<br />
0'<br />
) is live and t i<br />
⊆{ p|<br />
p∈P'<br />
∧(<br />
M'(<br />
p)<br />
> 0)}<br />
,<br />
then ( N , M<br />
0)<br />
and ( Nt , M t 0)<br />
are live.<br />
Theorem 4.3 Suppose that N , M ) is obtained from<br />
(<br />
0<br />
(<br />
0<br />
N ', M ')<br />
by t-subnet reduction operation. If<br />
( N , M<br />
0)<br />
and ( Nt , M t 0)<br />
are live, then ( N ', M<br />
0'<br />
) is<br />
live.<br />
Ⅴ. APPLICATIONS<br />
In this section we apply results of Section Ⅲ and<br />
Section Ⅳ to reduce a flexible manufacturing system.<br />
The manufacturing system consists of one workstation<br />
WS for assembly work and two machining centers for<br />
machining. Machining center_1 and WS share robot R<br />
1<br />
.<br />
Machining center_2 and WS share robot R<br />
2<br />
. The<br />
system runs as follows:<br />
In the machining center_1, the intermediate parts are<br />
machined by machine M<br />
1<br />
. Each part is fixtured to a<br />
pallet and loaded into the machine M<br />
1<br />
by robot R<br />
1<br />
.<br />
After processing, robot R 1<br />
unloads the final product,<br />
defixtures it and returns the fixture to M<br />
1<br />
.<br />
In the machining center_2, parts are machined first by<br />
machine M<br />
3<br />
and then by machine M<br />
4<br />
. Each part is<br />
automatically fixtured to a pallet and loaded into the<br />
machine. After processing, robot R<br />
2<br />
unloads the<br />
intermediate part from M<br />
3<br />
into buffer B. At machining<br />
station M<br />
4<br />
, intermediate parts are automatically loaded<br />
into M<br />
4<br />
and processed. When M<br />
4<br />
finishes processing a<br />
part, the robot R<br />
2<br />
unloads the final product, defixtures it<br />
and returns the fixture to M<br />
3<br />
.<br />
When workstation WS is ready to execute the<br />
assembly task, it requests robot R 1<br />
, robot R 2<br />
and<br />
machine M<br />
2<br />
and acquires them if they are available.<br />
When the workstation starts an assembly task, it cannot<br />
be interrupted until it is completed. When WS completes,<br />
it releases the robot R1<br />
and robot R<br />
2<br />
.<br />
Firstly, we give the Petri-net based model of the<br />
manufacturing system. Secondly, a reduced net system is<br />
obtained by p-subnet reduction method and t-subnet<br />
reduction method. Thirdly, we will analysis property<br />
preservation of the reduced net system.<br />
The Petri-net based model ( N , M<br />
0)<br />
of the original<br />
manufacturing system is illustrated in Fig.5.1.<br />
p 11<br />
t 11<br />
p 12<br />
p 13<br />
t 12<br />
p 14<br />
p 15<br />
t 13<br />
p 16<br />
p 17<br />
t 14<br />
t 31<br />
p r11<br />
t r11 t r12 p 33<br />
t<br />
p 32<br />
r13<br />
p 31<br />
p 34 p32<br />
p r12<br />
t 33<br />
t r13<br />
t r14<br />
p r14<br />
t 34<br />
p 35<br />
p 21<br />
t 21<br />
p<br />
p 23<br />
22<br />
p r21<br />
t 22<br />
t r21 p 24<br />
p r22<br />
p r23 t 23<br />
p 25<br />
p 26<br />
t 24<br />
t r22<br />
p 27<br />
t 25<br />
p r24<br />
t 26<br />
Fig. 5.1 The original net system<br />
p 28<br />
p 29<br />
252