06.03.2013 Views

Artificial Intelligence and Soft Computing: Behavioral ... - Arteimi.info

Artificial Intelligence and Soft Computing: Behavioral ... - Arteimi.info

Artificial Intelligence and Soft Computing: Behavioral ... - Arteimi.info

SHOW MORE
SHOW LESS

Create successful ePaper yourself

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

Definition 10.4: A transition to place connectivity (TPC) matrix P is a<br />

binary matrix whose element pij =1 if pi ∈ O (trj), otherwise pi j = 0. With n<br />

places <strong>and</strong> m transitions in the FPN, the P matrix is of (n × m) dimension.<br />

Since the binary product (AND-OR composition) (P o Q) represents mapping<br />

from places to their immediately reachable places, therefore, the presence of a<br />

'one' in the matrix M1 = (P o Q) at position (j, i) represents that pj ∈ IRS(pi).<br />

Analogously, a 'one' at position (j, i) in matrix Mr = (P o Q) r for positive<br />

integer r represents pj ∈ IRS r (pi), i.e., pj is reachable from pi with a degree of<br />

reachability r.<br />

Definition 10.5: If an element mij of matrix Mk = (P o Q) k is unity for<br />

positive integer k, then pi <strong>and</strong> pj are called associative places with respect<br />

to mij .<br />

Theorem 10.1: If the diagonal elements mii of the matrix Mk = (P o Q) k are<br />

unity, then the associative places pi for all i lie on cycles through k<br />

transitions in each cycle.<br />

Proof: The proof is presented in Appendix C.<br />

Corollary 1: In a purely cyclic FPN [34], where all transitions <strong>and</strong> places lie<br />

on a cycle, Mk = (P o Q) k = I, where k is the number of transitions (places) in<br />

the FPN.<br />

For identifying cycles in a FPN, the matrix Mk =(P o Q ) k for k = 1 to m<br />

is to be computed, where m is the number of transitions in the network. Then<br />

by theorem 10.1, the associative places corresponding to the diagonal<br />

elements of Mk will lie on a cycle with k transitions on each cycle. However,<br />

if more than k number of diagonal elements are unity, then places lying on a<br />

cycle are to be identified by finding immediate reachability of places on the<br />

cycle using M1 = (P o Q) matrix.<br />

The algorithm for cycle-detection consists of several procedures.<br />

Procedure Find-places-on-cycle determines the set of places Sk that lie on<br />

cycles with k transitions. Procedure Find-IRS-places saves in Lk the<br />

connectivity between pairs of immediately reachable set of places, lying on<br />

cycles with k transitions. Procedure Find-connected-places-on-cycles<br />

determines the list of places ordered according to their immediate reachability<br />

on cycles with k transitions <strong>and</strong> saves them in Newlistk.. Procedure Puttransitions<br />

positions appropriate transitions in the list of places in Newlistk,<br />

so that places preceding <strong>and</strong> following a transition in the modified list<br />

Finallistk are its input <strong>and</strong> output places on a cycle with k-transitions. The<br />

variable 'cycles' in procedure Cycle-detection denotes the list of cycles.

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

Saved successfully!

Ooh no, something went wrong!