Supervisory control of discrete event systems under asynchronous spiking neuron P systems

Asynchronous spiking neural P systems with multiple channels (SNP systems) operate in an asynchronous pattern, in which the firing of enabled rules is not obligatory. Asynchronous models are one of the most potential research carriers of supervisory control. However, a systematic understanding and process of leveraging SNPs to address control issues in discrete event systems are still lacking. This paper extends a type of SNP formalization, called vector-SNP, to explore the system's intrinsic stability properties without configuration attention by introducing two classes of structural feature vectors: N invariants and R-invariants. Then, a typical style of specifications, namely neuron mutual exclusion inequality constraints (NMEICS), is advanced for supervisory control using SNPs. An NMEIC is enforced by explicitly adding a supervisor (group of observation and control monitors represented by SNPs) to an SNP system. On top of that, the supervisor construct algorithm that utilizes the linear algebraic calculations is established to obtain the supervisors for pre-designed control constraints. Finally, we evaluate the proposed supervisory control theory using SNPs on an automated manufacturing system. Experimental results achieve maximally permissive controlled behaviors that reserve all feasible legal states. (c) 2022 Elsevier Inc. All rights reserved.

Automated summary

This paper extends the formalization of vector-SNP by introducing feature vectors and proposes neuron mutual exclusion inequality constraints (NMEICS) for supervisory control. A supervisor construct algorithm based on linear algebraic calculations is established. Experimental results show successful implementation on an automated manufacturing system.