Robust Deadlock Control for Automated Manufacturing Systems Based on the Max-Controllability of Siphons

Gaiyun Liu,Yufeng Chen,Yuting Liu,Zhiwu Li,Naiqi Wu,Lingchun Zhang

doi:10.1109/access.2019.2924021

Abstract

There have been a variety of deadlock control strategies proposed for automated manufacturing systems (AMSs) without taking unreliable resources into account in the framework of Petri nets. However, in addition to deadlocks, resource malfunction problems may also arise to make a system collapse in the real world. This paper develops a new deadlock control method for a generalized system of simple sequential processes with resources (GS 3 PR), where a class of unreliable GS 3 PR (U-GS 3 PR) is considered. Recovery subnets are used to describe an unreliable resource failure and its recovery process in a Petri net model. First, we compute all the strict minimal siphons (SMSs) in a GS 3 PR net. Second, with the concept of max-controllability of siphons, a monitor is added to each SMS. However, the net loses liveness when unreliable resources are taken into consideration with recovery nets being added for their operation places. Then, by using the concept of constraint set, we add related arcs between the transitions in recovery subnets and original monitors. Finally, a robust deadlock controller is designed for the AMSs according to the max-controllability of siphons.

Highlights

As a computer-controlled system, an automated manufacturing system (AMS) can efficiently produce small and medium batch of products, which is composed of numerical control machines
This paper addresses the issue of designing a robust deadlock control strategy for an AMS in the framework of generalized S3PR (GS3PR) with an unreliable resource based on siphons instead of reachability graphs
GS3PR WITH AN UNRELIABLE RESOURCE Considering an AMS with an unreliable resource type, we model the behavior of resource failure and recovery by Petri nets (PNs)

Summary

INTRODUCTION

As a computer-controlled system, an automated manufacturing system (AMS) can efficiently produce small and medium batch of products, which is composed of numerical control machines. The robust supervisors derived in that paper allow that there are multiple types of unreliable resources in an S3PR Theoretically, designing such a policy consumes an unacceptable long computation time if a PN contains a large number of SMSs and its reachability space is huge. This paper addresses the issue of designing a robust deadlock control strategy for an AMS in the framework of GS3PR with an unreliable resource based on siphons instead of reachability graphs. It is well-known that the computation of siphons in a GS3PR is the same as that in an S3PR according to the definition of siphons. The basics of PNs and the definition of GS3PR are given in Appendix

DEADLOCK CONTROL FOR GS3PR BASED ON

11: Output ‘‘Undecided’’

CONCLUSION