On Computation Reduction of Liveness-Enforcing Supervisors

• Published in: IEEE access Vol. 5; pp. 14775 - 14786
• Main Authors: Wang, Chao, Wang, Anrong, Chen, Yufeng, Al-Ahmari, Abdulrahman, Li, Zhiwu
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2017; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Computational efficiency; Computational modeling; Computing costs; deadlock prevention; flexible manufacturing system; Flexible manufacturing systems; Frequency modulation; Integer programming; Linear programming; Manufacturing systems; optimal liveness-enforcing supervisor; Petri net; Petri nets; Resource allocation; Resource management; Supervisors; System recovery
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2017.2728080

It is well recognized that the computation of an optimal liveness-enforcing supervisor (optimal supervisor for short) for a sequential resource allocation system is intractable due to the nature of the problem itself, since an integer linear programming model has to be formulated to find its solution. In this paper, a novel vector covering approach is developed to reduce the computational cost of designing maximally permissive supervisors for flexible manufacturing systems (FMSs), by reducing the number of operation places and legal markings (LMs) that need to be considered. A vector covering approach is used to find the minimal covered set of first-met bad markings (FBMs). Then, a novel vector covering approach is proposed to reduce the set of LMs to a smaller covering set. The proposed method can reduce the number of both variables and constraints in the integer linear programming problems (ILPPs). It raises the efficiency of designing optimal supervisors by solving ILPPs.