Coordination control of discrete-event systems revisited

Komenda, Jan; Masopust, Tomáš; Schuppen, Jan H. van

doi:10.1007/s10626-013-0179-x

Cited by 35 publications

(50 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…If only one of the two relationships (13) or (14) holds, then . It means that the projection channels allow passing all the events from their reference events set.…”

Section: Bmentioning

confidence: 99%

“…Relative coobservability is closed under set union and there exists the supremal coobservable sublanguage of a given language. In the bottomup approach, a coordinator is constructed to remove the conflict between decentralized supervisors [13,14]. In the topdown approach, distributed supervisory control synthesized by supervisor localization, guarantees the non-conflicting between local controllers [15].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Distributed supervisory control with partial observation

Saeidi

Afzalian

Gharavian

2016

2016 4th International Conference on Control, Instrumentation, and Automation (ICCIA)

View full text Add to dashboard Cite

Distributed supervisory control is a method to synthesize local controllers without any restriction on observation of the plant. Since the set algebra preserves relationships between variables and provides a simpler way to develop theories, the observation properties are redefined and corresponding propositions are proved by set algebra, in this paper. Moreover, we define new observation properties, i.e. local normality and local relative observability for the local controllers which constructed by any localization algorithm that preserves control equivalent to monolithic supervisor w.r.t. the plant. These properties enable us to investigate the observation problem from monolithic to distributed supervisory control. Furthermore, observation equivalence property is defined corresponding to the control equivalence in distributed supervisory control with partial observation. It is proved that, providing observation equivalence of the local controllers to the monolithic supervisor w.r.t. the plant, the control equivalence is satisfied, if and only if the intersection of local events sets is a subset or equal to the global observable events set.

show abstract

“…If only one of the two relationships (13) or (14) holds, then . It means that the projection channels allow passing all the events from their reference events set.…”

Section: Bmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Distributed supervisory control with partial observation

Saeidi

Afzalian

Gharavian

2016

2016 4th International Conference on Control, Instrumentation, and Automation (ICCIA)

View full text Add to dashboard Cite

show abstract

“…We briefly recall the original coordination control framework developed in [17]. The notation is that of [17].…”

Section: A Original Coordination Control Frameworkmentioning

confidence: 99%

“…The main existential result follows. Theorem 2 ( [17]): Consider the setting of Problem 2. There exist nonblocking supervisors S 1 , S 2 , and S k such that…”

Section: A Original Coordination Control Frameworkmentioning

confidence: 99%

On a distributed computation of supervisors in modular supervisory control

Komenda

Masopust

Schuppen

2015

2015 International Conference on Complex Systems Engineering (ICCSE)

Self Cite

View full text Add to dashboard Cite

In this paper, we discuss a supervisory control problem of modular discrete-event systems that allows for a distributed computation of supervisors. We provide a characterization and an algorithm to compute the supervisors. If the specification does not satisfy the properties, we make use of a relaxation of coordination control to compute a sublanguage of the specification for which the supervisors can be computed in a distributed way. I. INTRODUCTION AND MOTIVATIONWe investigate distributed supervisory control of concurrent discrete-event systems. Supervisory control theory of discrete-event systems modeled as finite automata was introduced by Ramadge and Wonham [24] and studied by many others. It aims to guarantee that the control specifications consisting of safety and of nonblockingness are satisfied in the controlled (closed-loop) system. Safety means that the language of the closed-loop system is included in a prescribed specification, and nonblockingness means that all controlled behaviors can always be completed to a marked controlled behavior. Supervisory control is realized by a supervisor that runs in parallel with the system and imposes the specification by disabling, at each state, some of the controllable events in a feedback manner. Since only controllable specifications can be achieved, one of the key issues is the computation of the supremal controllable sublanguage of the specification, from which the supervisor can be constructed.Supervisory control theory is well developed for monolithic systems, i.e., systems where the plant is modeled as a single generator. However, most of the current complex engineering systems can be abstracted as a composition of many components. Systems that model technological systems typically consist of small generators that communicate with each other in a synchronous [5] or asynchronous [7] way. Such systems are often called modular (or concurrent or distributed) discrete-event systems. It is known that to compute the overall monolithic plant for such a system can be unrealistic because the number of states of a modular system grows exponentially with respect to the number of components. This limits the applicability of the monolithic supervisory control synthesis to relatively small systems. On the other hand, the purely decentralized control consisting of an independent construction of a supervisor for each subsystem is only applicable for a local (decomposable) specification.Specifically, let G 1 and G 2 be two systems over the respective alphabets Σ 1 and Σ 2 modeled as finite generators forming the overall plant G 1 G 2 , the computation of which we want to avoid, and let K ⊆ L m (G 1 G 2 ) denote a specification. There exists a monolithic supervisor for the monolithic plant G 1 G 2 if and only if the specification is controllable with respect to the plant. However, to avoid the computation of the monolithic plant G 1 G 2 , the naive approach to compute supervisors for each subsystem separately does not work in general. To demonstrate this, assume tha...

show abstract

“…It turns out that many problems tractable in the monolithic setting become intractable in the modular setting due to the state-space exploration issue. There are many results in the literature on finding tractable solutions for problems in the context of modular DES [5,7,8,11,12,18,20].…”

Section: Introductionmentioning

confidence: 99%

Complexity of detectability, opacity and A-diagnosability for modular discrete event systems

Masopust

Yin

2019

Automatica

Self Cite

View full text Add to dashboard Cite

We study the complexity of deciding whether a modular discrete event system is detectable (resp. opaque, A-diagnosable). Detectability arises in the state estimation of discrete event systems, opacity is related to the privacy and security analysis, and A-diagnosability appears in the fault diagnosis of stochastic discrete event systems. Previously, deciding weak detectability (opacity, A-diagnosability) for monolithic systems was shown to be PSPACE-complete. In this paper, we study the complexity of deciding weak detectability (opacity, A-diagnosability) for modular systems. We show that the complexities of these problems are significantly worse than in the monolithic case. Namely, we show that deciding modular weak detectability (opacity, A-diagnosability) is EXPSPACE-complete. We further discuss a special case where all unobservable events are private, and show that in this case the problems are PSPACE-complete. Consequently, if the systems are all fully observable, then deciding weak detectability (opacity) for modular systems is PSPACE-complete.

show abstract

Coordination control of discrete-event systems revisited

Cited by 35 publications

References 33 publications

Distributed supervisory control with partial observation

Distributed supervisory control with partial observation

On a distributed computation of supervisors in modular supervisory control

Complexity of detectability, opacity and A-diagnosability for modular discrete event systems

Contact Info

Product

Resources

About