PETRI NETWORKS
| Collection | International Encyclopedia of Systems and Cybernetics |
|---|---|
| Year | 2004 |
| Vol. (num.) | 2(2) |
| ID | ◀ 2539 ▶ |
| Object type | General information, Methodology or model |
- “An oriented graph used to describe the behavior of asynchronic sequential automata ” (A. LEROY and J.P. SIGNORET, 1992, p.49)
PETRI networks were introduced by Dr. PETRI in Germany in 1962. They are still in a stage of further development.
A. PETRI network is made of “places”or \term“{positions} ”which represent the static part of the network.
Tokens are moved from places to places in accordance with some transition rules which must be specified. Thus the network evolves from state to state by successive steps. During its evolution the PETRI network covers progressively all the possible states of the system under study.
It combines a structuralrepresentation of the modelized automata and its possibilities oftransformations, based on the transitions rules . These can bedeterministic with specifieddelays, orrandom ones, representing in such casesphenomena whose occurrence cannot be precisely forecasted.
The authors explain as follows the possible uses of PETRI nets, which sequentially follow the different successive states of the modelized system:
- analyse in a detailed way its sequential behavior
- identify the different states in view of creating a markovian graph of states
- identify the non-accessible states
- find the lock-ins , delays , feedbacks , etc
- discover bifurcations or different possible transitions (Ibid, p. 54)
PETRI networks can be used to modelize projected physical or mechanical systems and discover potential failures, risks or incompatibilities. (Adapted from A. LEROY and J.P. SIGNORET, 1992)