THE MINIMAL PATH UPPER BOUND FOR THE MOMENTS OF A RELIABILITY FUNCTION

We consider a binary system with binary components, conditionally inde pendent components given the component reliabilities, and random, posi tively dependent (associated) component reliabilities. An upper bound is derived for the moments of the (random) reliability function, corre sponding to the well-known minimal path upper bound for the non-random case. This extends a result of Natvig & Eide (1987). In addition we u se the inequality to derive a new upper bound for the second moment of the reliability function, based on minimal cut sets.