10.01.2013 Views

ComputerAided_Design_Engineering_amp_Manufactur.pdf

ComputerAided_Design_Engineering_amp_Manufactur.pdf

ComputerAided_Design_Engineering_amp_Manufactur.pdf

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

needs to show that after each full synthesis step using Rule TT.4 or PP.2, the resulting remains so.<br />

�� that does not include any node in the NP, it will fire in exactly the same manner as that in N and<br />

leads to the same reachable marking. Hence this case need not be considered.<br />

Reversible: We need to show that a � exists such that . There are two cases:<br />

(a) and (b) .<br />

(a) Two cases if � (1) does or (2) does not include any t � . For case (2), the subnets that involve<br />

� are exactly those in . Hence, the � exists. For case (1), Lemma 5 dictates that a .<br />

Let , then [ ], where . The problem then<br />

reduces to case (2) which has been proved.<br />

(b) , which implies a partial . By Lemma 5, all sbsequent firing sequences, if long enough,<br />

can complete the rest of to restore . The case then is similar to case (a) and can be proved<br />

similarly.<br />

Live: Assume to the contrary, then � , t � ; some of its input places, which<br />

are mutually concurrent, can never get tokens. But this is impossible by the equivalence between temporal<br />

and structural relationship (which ensures that all input places of t are possible to get tokens) and the<br />

requirement that the NPs in Rule TT.4.2 and PP.2.2 should be virtual (which ensures that the above<br />

possibility should be definite) by preventing uncoordinated movement of tokens in a set of concurrent<br />

places. When a token enters a place with more than one output transition, it may fire any of its output<br />

transitions and thus it may result in uncoordinated movement of tokens.<br />

That is, the two tokens generated by firing a transition may move to the input places of two exclusive<br />

transitions, respectively. Since they are mutually exclusive, none of them will be able to be fired; the N hence<br />

is not live. The virtual PSP generations in Rules TT. 4.2 and PP.2.2 prevent such uncoordinated movements.<br />

One of the above two tokens moves in ; the other in the NP. Uncoordinated movements of these<br />

two tokens intrude the marking behavior of ; this is impossible by Lemma 5, which states that<br />

guideline (2) is satisfied.<br />

Note that Guideline (1) can only be revoked when the a has input places from the NP. Then, by<br />

Rule TT.2, the NP will have a token in its initial marking, which allows to be potentially able to be<br />

fired from .<br />

Thus both guidelines are satisfied. Based on this, we can show that the t is potentially able to be fired.<br />

If (1) , then by Lemma 4, the presence of NPs does not change the markings of ; hence,<br />

if t is not in the NP, t is potentially able to be fired. If t is in the NP, then by Observation 1 and Guideline<br />

(1), t is also potentially able to be fired. Now if (2) , then also by Lemma 4 there exists a<br />

subsequent firing sequence to drive back to and then is a reachable marking in .<br />

The case then degenerates to case (1).<br />

1a<br />

1a w<br />

1a w<br />

( M� , M ) [� � ( M0 , M0 )<br />

M w w<br />

� M0 M w w<br />

�M0 T w<br />

N 1a<br />

� w<br />

� �<br />

� �c� w 1a w<br />

� �d ( M� , M0 ) �1� w 1a w<br />

1a 1a<br />

� ( M� , M0 ) M� �R<br />

M w w<br />

�M0 � w<br />

� w<br />

w<br />

M0 � a R 2a<br />

1a w<br />

( M� ,M ) � T 2<br />

N 1a<br />

N 1a<br />

tg tg 2a<br />

M0 M w w<br />

� M0 N 1a<br />

M w w<br />

�M0 M w<br />

w<br />

M0 M 1/2<br />

M 1a<br />

Theorem 4: Any PN resulting from a singular application of Rule TT.3 to an synthesized from<br />

a basic process is well-behaved.<br />

Proof: Lemma 5 does not hold for Rule TT.3.1 because Lemma 3 does not either. This is because<br />

may fire infinitely often relative to tj . Now Rule TT.3.2 prevents such infinite firings of tg with respect<br />

to tj and forces dgj � 1. Lemma 3, and hence Lemma 5 also hold for Rule TT.3. The rest of proof follows<br />

that for Rule TT.4 in Theorem 3.<br />

Theorem 5: Any PNs resulting from the applications of the TT and PP rules to a basic process are<br />

bounded, live, reversible, and marking monotonic.<br />

Proof: Prove by induction. It is easy to see that a basic process is live, bounded, and reversible. Then,<br />

assuming the N from the (N-1)th synthesis steps is bounded, live, and reversible, the proof needs to show<br />

that after each full synthesis step using a certain TT or PP rule, the resulting N remains so. The<br />

correctness of this is established by Theorems 1–4. Note that we have been assuming that there is only<br />

one token in every home place. Now prove the property of marking monotonic by showing that is<br />

well-behaved.<br />

1a<br />

N b<br />

N 1a<br />

N 2a<br />

t g

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!