An Inverse Method for Parametric Timed Automata

André, Étienne; Chatain, Thomas; Fribourg, Laurent; Encrenaz, Emmanuelle

doi:10.1142/s0129054109006905

Cited by 75 publications

(72 citation statements)

References 15 publications

(10 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The tool Imitator (Inverse Method for Inferring Time AbstracT behaviOR) implements the algorithm InverseMethod , described in [4]. We assume given a system modeled by a PTA A.…”

Section: Principle Of Imitatormentioning

confidence: 99%

“…This is written A[π] ≡ TA A[π 0 ]. More formally, the algorithm InverseMethod solves the following inverse problem [4] for acyclic systems (i.e., with only finite traces) by computing a constraint K 0 such that :…”

Section: Principle Of Imitatormentioning

confidence: 99%

“…In order to handle cyclic examples, one modifies the algorithm by replacing, in the IF condition, Post A(K) (S) = ∅ by Post A(K) (S) ⊆ S. In that case, we ensure termination more often (see [4]). However, we do not guarantee any longer the identity of traces, but only the identity of reachable locations.…”

Section: General Structurementioning

confidence: 99%

See 2 more Smart Citations

IMITATOR: A Tool for Synthesizing Constraints on Timing Bounds of Timed Automata

André

2009

Theoretical Aspects of Computing - ICTAC 2009

Self Cite

View full text Add to dashboard Cite

Abstract. We present here Imitator, a tool for synthesizing constraints on timing bounds (seen as parameters) in the framework of timed automata. Unlike classical synthesis methods, we take advantage of a given reference valuation of the parameters for which the system is known to behave properly. Our aim is to generate a constraint such that, under any valuation satisfying this constraint, the system is guaranteed to behave, in terms of alternating sequences of locations and actions, as under the reference valuation. This is useful for safely relaxing some values of the reference valuation, and optimizing timing bounds of the system. We have successfully applied our tool to various examples of asynchronous circuits and protocols. ContextTimed automata [1] are finite control automata equipped with clocks, which are real-valued variables which increase uniformly. This model is useful for reasoning about real-time systems, because one can specify quantitatively the interval of time during which the transitions can occur, using timing bounds. However, the behavior of a system is very sensitive to the values of these bounds, and it is rather difficult to find their correct values. It is therefore interesting to reason parametrically, by considering that these bounds are unknown constants, or parameters, and try to synthesize a constraint (i.e., a conjunction of linear inequalities) on these parameters which will guarantee a correct behavior of the system. Such automata are called parametric timed automata (PTA) [2,11].The synthesis of constraints for PTA has been mainly done by supposing given a set of "bad states" (see, e.g., [8,9]). The goal is to find a set of parameters for which the considered timed automaton does not reach any of these bad states. We call such a method a bad-state oriented method. By contrast, we present in this paper a tool based on a good-state oriented method. Principle of ImitatorThe tool Imitator (Inverse Method for Inferring Time AbstracT behaviOR) implements the algorithm InverseMethod , described in [4]. We assume given a

show abstract

“…The tool Imitator (Inverse Method for Inferring Time AbstracT behaviOR) implements the algorithm InverseMethod , described in [4]. We assume given a system modeled by a PTA A.…”

Section: Principle Of Imitatormentioning

confidence: 99%

Section: Principle Of Imitatormentioning

confidence: 99%

See 1 more Smart Citation

IMITATOR: A Tool for Synthesizing Constraints on Timing Bounds of Timed Automata

André

2009

Theoretical Aspects of Computing - ICTAC 2009

Self Cite

View full text Add to dashboard Cite

show abstract

“…We say that A behaves well if it satisfies a certain set of properties. We are interested here in properties that are invariant for automata having the same set of traces (alternating sequences of locations and actions, i.e., time-abstract runs) [3]. This is in particular the case of linear-time properties [7].…”

Section: Introductionmentioning

confidence: 99%

“…We generate a constraint on the parameters ("tile") for each integer point located within a given rectangle V 0 . Such a tile is called "behavioral tile" because A behaves similarly under any parameter valuation corresponding to a point of the tile: the sets of traces coincide [3]. This allows us to decompose the parametric space into behavioral tiles.…”

Section: Introductionmentioning

confidence: 99%