A simulation study of estimates of system availability.

Abstract : A system which is either working or under repair is modeled as an alternating renewal process. Data is in the form of a finite number of independent lifetimes and repair times. Several semi-parametric estimators of the probability that the system is up at a finite time t are studied via simulation. The estimators are based on an exponential approximation to the true system availability at time t and use empirical Laplace transforms of the lifetimes and repair times. Keywords: Bootstrap confidence intervals, Tables data, Random variables, Theses.