Implementation of generalized mutual exclusion constraints using critical places and marking estimation
Generalized mutual exclusion constraints (GMECs) are a class of state specifications on Petri nets (PNs). They are generally enforced on the nets by a simple control structure called control places (monitors). Unfortunately, this conventional procedure is implemented in an offline and monolithic man...
Saved in:
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/159948 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Generalized mutual exclusion constraints (GMECs) are a class of state specifications on Petri nets (PNs). They are generally enforced on the nets by a simple control structure called control places (monitors). Unfortunately, this conventional procedure is implemented in an offline and monolithic manner, which suffers from computational difficulty and is arduous for the control of a system in real time. Additionally, the flexibility and fault tolerance of such a method are somewhat unacceptable, and the method suffers from a lack of adaptability to net variations incurred by system reconfiguration, communication failure, and constraint integration. This article aims to enforce GMECs for a live PN model by using critical places and marking estimation. First, we define GMECs on some so-called critical places, such that the satisfaction of the GMECs can be determined by only monitoring the number of tokens in their corresponding critical places during runtime. Then, an efficient and effective control strategy is developed such that the controllers forbid all those transition firings that lead to the violated GMECs based on the estimated markings of critical places derived by observers from a resource perspective rather than exploring the markings of the entire system. Finally, we present procedures to deal with decision deadlocks, which are induced by one-sided decisions made by some controllers and may prevent all enabled transitions from firing. Global GMECs are always implemented through the local observation and control of processes without knowing an extra information. |
|---|