Seventh International Conference on Simulation Tools and Techniques

Research Article

A data type for discretized time representation in DEVS

Download629 downloads
  • @INPROCEEDINGS{10.4108/icst.simutools.2014.254803,
        author={Damian Vicino and Olivier Dalle and Gabriel Wainer},
        title={A data type for discretized time representation in DEVS},
        proceedings={Seventh International Conference on Simulation Tools and Techniques},
        publisher={ICST},
        proceedings_a={SIMUTOOLS},
        year={2014},
        month={8},
        keywords={simulation data type time devs},
        doi={10.4108/icst.simutools.2014.254803}
    }
    
  • Damian Vicino
    Olivier Dalle
    Gabriel Wainer
    Year: 2014
    A data type for discretized time representation in DEVS
    SIMUTOOLS
    ICST
    DOI: 10.4108/icst.simutools.2014.254803
Damian Vicino1,*, Olivier Dalle1, Gabriel Wainer2
  • 1: Université Nice Sophia Antipolis
  • 2: Carleton University
*Contact email: damian.vicino@inria.fr

Abstract

This paper addresses the problems related to data types used for time representation in DEVS, a formalism for the specification and simulation of discrete-event systems. When evaluating a DEVS simulation model into an actual computer simulation program, a data type is required to hold the virtual time of the simulation and the time elapsed in the model of the simulated system. We review the commonly data types used, and discuss the problems that each of them induce. In the case of floating point we show how, under certain conditions, the simulation can break causality relations, treat simultaneous events as non simultaneous or treat non simultaneous events as simultaneous. In the case of integers using fixed unit we list a number of problems arising when composing models operating at different timescales. In the case of structures that combine several fields, we show that, at the cost of a lower performance, most of the previous problems can be avoided, although not totally. Finally, we describe an alternative representation data type we developed to cope with the data type problems.