Quantifying Fidelity for Timed Transition Systems - LAAS - Laboratoire d'Analyse et d'Architecture des Systèmes Access content directly
Conference Papers Year : 2016

Quantifying Fidelity for Timed Transition Systems

Sangeeth Saagar Ponnusamy
  • Function : Author
  • PersonId : 980153
Vincent Albert
  • Function : Author
  • PersonId : 865659
Patrice Thebault
  • Function : Author
  • PersonId : 980063


The paper addresses one of the fundamental questions in using simulation as a means for system verification and validation, namely, how far the simulation model represents the transition timings of the real system. A formal quantification of this difference in transition timings of a simulation model with respect to the system specification is presented based on game theoretic distance notions from literature. In this two player game, simulation model tries to mimic the system's transitions and incurs a distance if it fails to match the timing of the transition. Fidelity of simulation model is presented through this distance notion based on the quantitative simulation relations and timed simulation game. This game between two timed transition systems is modeled in petri-net formalism and a quantitative reachability graph is generated using TINA tool embedded in ProDEVS simulation platform to explore all such player strategies. The resulting exhaustive exploration yields a global fidelity distribution of the simulation model in terms of transition timings which could be analysed in ProDEVS to gain further insight into the simulation model behaviour with respect to the system model. The approach is demonstrated on a buffer system modeling case study to validate a processor through simulation.
Fichier principal
Vignette du fichier
Quantifying Fidelity for Timed Transition Systems.pdf (395.19 Ko) Télécharger le fichier

Dates and versions

hal-01912574 , version 1 (18-01-2019)


  • HAL Id : hal-01912574 , version 1


Sangeeth Saagar Ponnusamy, Vincent Albert, Patrice Thebault. Quantifying Fidelity for Timed Transition Systems. 6th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH 2016), Jul 2016, Lisbonne, Portugal. ⟨hal-01912574⟩
25 View
10 Download


Gmail Facebook X LinkedIn More