EAI Endorsed Transactions on Industrial Networks and Intelligent Systems 16(9): e3

Research Article

Using Finite Forkable DEVS for Decision-Making Based on Time Measured with Uncertainty

Download20 downloads
  • @ARTICLE{10.4108/eai.24-8-2015.2261152,
        author={Damian Vicino and Olivier Dalle and Gabriel Wainer},
        title={Using Finite Forkable DEVS for Decision-Making Based on Time Measured with Uncertainty},
        journal={EAI Endorsed Transactions on Industrial Networks and Intelligent Systems},
        volume={16},
        number={9},
        publisher={ACM},
        journal_a={INIS},
        year={2015},
        month={8},
        keywords={devs, uncertainty, metrology, time interval},
        doi={10.4108/eai.24-8-2015.2261152}
    }
    
  • Damian Vicino
    Olivier Dalle
    Gabriel Wainer
    Year: 2015
    Using Finite Forkable DEVS for Decision-Making Based on Time Measured with Uncertainty
    INIS
    EAI
    DOI: 10.4108/eai.24-8-2015.2261152
Damian Vicino1, Olivier Dalle,*, Gabriel Wainer2
  • 1: Laboratoire I3S UMR CNRS 7172
  • 2: Carleton University
*Contact email: olivier.dalle@inria.fr

Abstract

The time-line in Discrete Event Simulation (DES) is a sequence of events defined in a numerable subset of R+. When it comes from an experimental measurement, the timing of these events has a limited precision. This precision is usually well-known and documented for each instruments and procedures used for collecting experimental datas. Therefore, these instruments and procedures produce measurement results expressed using values each associated with an uncertainty quantification, given by uncertainty intervals. Tools have been developed in Continuous Systems modeling for deriving the uncertainty intervals of the final results corresponding to the propagation of the uncertainty intervals being evaluated. These tools cannot be used in DES as they are defined, and no alternative tools that would apply to DES have been developed yet. In this paper, we propose simulation algorithms, based on the Discrete Event System Specification (DEVS) formalism, that can be used to simulate and obtain every possible output and state trajectories of simulations that receive input values with uncertainty quantification. Then, we present a subclass of DEVS models, called Finite Forkable DEVS (FF-DEVS), that can be simulated by the proposed algorithms. This subclass ensures that the simulation is forking only a finite number of processes for each simulation step. Finally, we discuss the simulation of a traffic light model and show the trajectories obtained when it is subject to input uncertainty.