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

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Note if the pair of nodes are both output (input) nodes of a prime start (end) node, then the corresponding<br />

prime end (start) node must be the same type as the prime start (end) node; i.e., they must both be<br />

transitions or both be places. For instance, t1 and t3 (p2 and p�2 ) are both output (input) nodes of<br />

p1( t6); their prime end is also a place (transition) p2 or pg ( t1 or t2 or t3) . Thus, in an SC, the two<br />

prime nodes of a pure pair must both be transitions or places.<br />

Definition: If �� , n1�� and n2��, then n1 ↔ n2 . If n1 � �1 and n2 � �2 , �1 � �2 ; the<br />

structural relationship between n1 and n2 is A12 .<br />

Independent, uncoordinated action may cause an N to be unbounded and nonlive. A single generation<br />

may thus necessitate that additional generations maintain coordination and the associated searching of<br />

additional ng and nj as in Rules TT.4 and PP.2 below. This searching motivates the following two<br />

definitions.<br />

Definition: A local exclusive set (LEX) of �i with respect to �k, Xik , is the maximal set of all PSPs that<br />

are exclusive to each other and are equal or exclusive to �i, but not to �k. That is, Xik � LEX ( �i,� k)<br />

� {�z��z � �i or �z��i, ¬ ( �z��k) , ��z1,� z2 � Xik,� z1 ��z2} .<br />

Definition: A local concurrent set (LCN) of �i with respect to �k, Cik, is the maximal set of all PSPs<br />

which are concurrent to each other and are equal or concurrent to �i, but not to �k, i.e., Cik � LCN<br />

( �i,� k)<br />

�{�z��z � �i or �z���i, ¬ ( �z||�k) , ��z1,� z2 � Cik,� z1 || �z2} .<br />

Ex<strong>amp</strong>les of LEX and LCN appear in Figures 8.5 and 8.6, and Rules TT.4 and PP.2, respectively, employ<br />

them.<br />

The definition of LEX (LCN) between two PSPs can be extended to the LEX (LCN) between two<br />

transitions (places). Thus LEX(ti, tk) [LCN ( pi,p k)<br />

] is a set of transitions (places), instead of PSPs. Let<br />

G(J)denote the set of all �g( �j) involved in a single application of the TT or PP rule. To avoid the<br />

unboundedness problem, it is necessary to have a new directed path from each PSP in Xgj to each PSP<br />

in Xjg . Both Xgj and Xjg can be determined from the T-Matrix even though dgj � �,<br />

which is,<br />

therefore, of no use to the synthesis.<br />

Since synchronic distance depends on the marking, the determination of synchronic distance between<br />

all pairs of transitions may take exponential time complexity. It is easier, however, to determine the<br />

synchronic distance due to the structure of the net; such a synchronic distance is called the structure<br />

synchronic distance d as defined below. Since we construct most new generations based on the net<br />

structure, rather than on the marking, of interest is the structure of synchronic distance, rather than the<br />

synchronic distance itself. We may find the structure synchronic distance of a synthesized net by examining<br />

the synchronic distance of the same net with each home place holding only one token; i.e., by<br />

making the net safe. However, not all such synchronic distances are useful to the synthesis. For instance,<br />

we will observe that when and d � 1, the information of is of no use and we<br />

define that . On the contrary, when d , the TT-generation from to may<br />

cause unbounded places in the TT-path. This problem cannot be detected by checking the T-Matrix only.<br />

The information of is indeed useful here, and the value of is assigned to be . Hence we<br />

have following definition:<br />

s<br />

( )<br />

tg � tj ( Xgj, Xjg) dgj � �<br />

s<br />

dgj � 1<br />

( Xgj, Xjg) � �<br />

tg tj s<br />

dgj � �<br />

dgj �<br />

Definition (Structure Synchronic Distance): The structure synchronic distance between transitions<br />

s<br />

s<br />

and tj is dgj � 1 if d ( Xgj, Xjg) is finite and dgj � � if d ( Xgj, Xjg) � �.<br />

8.3 The Synthesis Rules<br />

We first present possible types of generations followed by the definitions of the rules.<br />

In order for the rules to be complete, all possible generations must be considered. The types of<br />

generations depend on ng and nj in two factors: (1) whether they are transitions or places and (2) their<br />

structural relationship. For factor 1, there are four types, TT, PP, TP, and PT generations, defined as<br />

follows:<br />

t g

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

Saved successfully!

Ooh no, something went wrong!