Design of deadlock prevention supervisors using Petri nets

This paper proposes a methodology to synthesize supervisors for a class of sequential resource allocation system for flexible manufacturing systems. The type of Petri nets are called S3PR, where deadlocks are related to emptied siphons. In a former paper (Huang et al., IEEE Trans Syst Man Cybern, 20...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:International journal of advanced manufacturing technology 2007-12, Vol.35 (3-4), p.349-362
1. Verfasser: Huang, Yi-Sheng
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:This paper proposes a methodology to synthesize supervisors for a class of sequential resource allocation system for flexible manufacturing systems. The type of Petri nets are called S3PR, where deadlocks are related to emptied siphons. In a former paper (Huang et al., IEEE Trans Syst Man Cybern, 2007), a deadlock prevention policy was proposed based on Petri nets siphons for the type of Petri nets. Since all minimal siphons should be controlled, the deadlock prevention policy is very time-consuming when the system is large. In this research, a concept of the elementary siphon is used to reduce the number of control places. A new siphon-based policy of deadlock prevention for the type of Petri nets is presented. This policy consists of two main stages: The first stage, called elementary siphons control, adds control places to the original net model to prevent elementary siphons from being emptied. The second stage, called generalized siphons control, adds control places that adopt a conservative policy of controlling only the release of parts into the system are used. Compared to the existing approaches, the new deadlock prevention policy can obtain a structurally compact deadlock prevention supervisor by adding only a few control places. Finally, numerical experiments under reachable states illustrate that the proposed algorithm appears to be more permissive than the closely related approaches.
ISSN:0268-3768
1433-3015
DOI:10.1007/s00170-006-0708-y