New Petri Net Structure and Its Application to Optimal Supervisory Control: Interval Inhibitor Arcs


Chen Y., Li Z., Barkaoui K., UZAM M.

IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS, cilt.44, sa.10, ss.1384-1400, 2014 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 44 Sayı: 10
  • Basım Tarihi: 2014
  • Doi Numarası: 10.1109/tsmc.2014.2307284
  • Dergi Adı: IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.1384-1400
  • Anahtar Kelimeler: Deadlock, flexible manufacturing system, interval inhibitor arc, Petri net, supervisory control, FLEXIBLE MANUFACTURING SYSTEMS, DEADLOCK PREVENTION POLICY, LIVENESS-ENFORCING SUPERVISORS, RESOURCE-ALLOCATION SYSTEMS, DEPENDENT SIPHONS, REGIONS, CONTROLLABILITY, ELEMENTARY, AVOIDANCE, MODELS
  • Yozgat Bozok Üniversitesi Adresli: Hayır

Özet

This paper presents a new Petri net structure, namely, an interval inhibitor arc, and its application to the optimal supervisory control of Petri nets. An interval inhibitor arc is an arc from a place to a transition labeled with an integer interval. The transition is disabled by the place if the number of tokens in the place is between the labeled interval. The formal definition and the firing rules of Petri nets with interval inhibitor arcs are developed. Then, an optimal Petri net supervisor based on the interval inhibitor arcs is designed to prevent a system from reaching illegal markings. Two techniques are developed to simplify the supervisory structure by compressing the number of control places. The proposed approaches are general since they can be applied to any bounded Petri net models. A marking reduction approach is also introduced if they are applied to Petri net models of flexible manufacturing systems. Finally, a number of examples are provided to demonstrate the proposed approaches and the experimental results show that they can obtain optimal Petri net supervisors for some net models that cannot be optimally controlled by pure net supervisors. Furthermore, the obtained supervisor is structurally simple.