Using Simulation to Estimate the Reliability of Systems of
Dependent Components
Consider a system of components, each of which either works or is failed.
Assume a random environmental parameter V having a known distribution and
suppose that the components are conditionally independent, given V, with
component i working with probability pi. For a given structure we present a new
approach to using simulation to efficiently estimate the probability the system
works. We will also discuss a multi-state generalization where the state of a
component is, conditional on V, a Poisson random variable.