Structural Decomposition and Decentralized Control of Petri Nets

Document Type

Article

Publication Date

8-1-2018

Abstract

Control of a large-scale automated manufacturing system is an important and challenging issue. Its discrete event system model represented by Petri nets tends to become highly complicated in structure, especially when there exist uncontrollable or unobservable events. The existing approaches are nontrivial to design both efficient and maximally permissive supervisors to impose constraints on an overall system. In this paper, instead of considering the control problem from an overall system perspective, we intend to transform an overall control problem into the one designing multiple controllers in parallel, each of which is much simpler in structure. A Petri net structure is decomposed via integer linear programming or a polynomial decomposition method to obtain multiple state-machine subnets that constitute a decentralized system. A necessary and sufficient condition for preserving the equivalence in terms of states and behaviors between the overall system and its decentralized version is reported. Constraints representing control requirements are further converted and enforced in the respective subnets. Then, supervisors are generated via a generalized mutual exclusion constraint method. By considering the deviations between the subnet control and overall control, this paper formulates a communication mechanism to guarantee that the decentralized system runs in an appropriate manner. Finally, two examples are presented to demonstrate the proposed approach.

Identifier

85028833536 (Scopus)

Publication Title

IEEE Transactions on Systems Man and Cybernetics Systems

External Full Text Location

https://doi.org/10.1109/TSMC.2017.2703950

e-ISSN

21682232

ISSN

21682216

First Page

1360

Last Page

1369

Issue

8

Volume

48

Grant

12-ELE2506-02

Fund Ref

National Natural Science Foundation of China

This document is currently not available here.

Share

COinS