STATISTICAL SIMULATION OF RELIABILITY OF NETWORKS WITH EXPONENTIALLY DISTRIBUTED UNIT LIFETIMES

In this paper, there were deduced three new lifetime distributions of serial-parallel and parallel-serial networks, their distribution being approached by means of analytical and Monte-Carlo methods. The novelty of the distribution consists in the fact that the number of subnets is random, governed by the Poisson and Logarithmic distributions, the lifetimes of the units in each subnet being independent, identically, exponentially distributed random variables, the number of units in each subnet is the same constant integer number. It was shown that the most important theoretical characteristics of lifetime for such networks, as the mean value, the variance, the survival/reliability function, my be approximated, with desired accuracy, by the same corresponding characteristics, as the sample mean value, sample variance, empirical survival/reliability function simulated by Monte-Carlo methods. Results are illustrated tabularly and graphically for some concrete examples.