he research of analytical solutions for reliability assessment in dynamic context is not solved in the general case. Partial approaches exist in the case of particular hypothesis. The Monte Carlo simulation would be the only recourse, but there were no efficient tools allowing the simultaneous simulation of the discrete event evolution of the system and its continuous evolution, taking into account the probabilistic aspects. In this context, we have formally introduced the concept of stochastic hybrid automaton (SHA) with the goal of consider all the problems defining the concept of the dynamic reliability. The SHA has been implemented in the Scicos toolbox of the Scilab environment. Finally, we have defined the DyRelA tool with the aim to access to the assessment of the dependability parameters by a Monte Carlo simulation of a system modeled by a SHA. The goal of this article is to present an improved version of the DyRelA tool which will significantly reduce the simulation duration of the dependability assessment of a system in dynamic context.
Gabriel Antonio Pérez Castañeda, Nicolae Brinzei, Jean-François Aubry