A Matrix Vector Transition Net Implementation | Chapter 01 | Advances in Mathematics and Computer Science Vol. 1
Aims:
Classic Petri nets also known as place transition nets provide many interesting
and useful features for system modeling. They are however limited by the place
types that are used. A novel approach is presented in this work. A matrix
vector transition net model is created and is used to model complex system
behavior. This solution extends the modeling power of normal Petri nets.
Proposed
Solution: A traditional Petri net is modified to
create a matrix vector transition net (MVTN). The idea is to combine the ideas
from normal Petri net semantics with a matrix vector approach.
Implementing
the Matrix Vector Transition Net: Ordinary Petri net places are replaced
with matrices or vectors. The input and output arcs must have a specific
function matrix that determines firing. Firing and behavior remain conceptually
and functionally similar to that of a Petri net. It is possible to interchange
row and column vectors. The behavior of matrix transition nets must elicit
similar behavior to that of a place transition net. Instead of normal tokens,
matrix elements are used. The matrix vector type of structure increases the
modeling power, abstraction capacity and the complexity of the net.
Case
Study: To illustrate this work a toy case of
an abstract network structure containing processing elements is used to
illustrate the use of the matrix vector transition net structure.
Results
and Findings: The behavior of matrix transition nets
is shown to be similar in principle to that of a place transition net. However
instead of tokens, matrix elements are used. It is possible to construct a
symbolic marking graph or reachability graph for the system This type of
structure definitely increases the modeling power, abstraction capacity and the
complexity of the net. The matrix transition net could be useful for certain
types of communication system problems and complex system interfacing. Several other uses can be found for this
approach in both computing and modeling.
Author Details:
Tony Spiteri Staines
Department
of Computer Information Systems, University of Malta, MSIDA MSD 2080, Malta.
Comments
Post a Comment