An automated technique to support the verification and validation of simulation models

Abstract

Simulation modeling requires model validation and ver ification to ensure that computed results are worth being considered. While we cannot expect a magic solution to the general problem, automated techniques for particular aspects of validation and verification are feasible. In this paper, we propose a technique to deduce model properties automatically from simulation runs performed for verifica tion and validation and to use those properties for runtime monitoring during production runs. Properties are repre sented as formulas in linear temporal logic and are limited to functional properties. We demonstrate the applicability of the approach with using an extended version of a stochas tic Botnet model originally developed by Van Ruitenbeek and Sanders.