AN ALGORITHM OF ACCELERATED PROBABILISTIC MODELING FOR ESTIMATION OF THE RELIABILITY OF COMPLEX-SYSTEMS

An algorithm is set up for estimating the reliability of complex systems by the accelerated Monte-Carlo method with the modeling results presented in the form of the sample stratified relative to the number of failures of elements. It is proved that the estimate of the system reliability factor, which as given in the form of a weighted sum of the stratum-by-stratum estimates with analytically calculated weight factors, is consistent and unbiased. Compared to the direct method of probabilistic modeling, the suggested method permits reducing the size of the computer (-based) experiment by both excluding from the draw (game) certainly failure-free trials (trial outcomes) and unfailing (or unfailed) elements and applying analytical models invariant relative to the characteristics of an engineering system modeled and the adopted procedure of maintenance of the systems tinder study.