Bootstrap resampling techniques in bootstrap information criteria for reliability analysis

Abstract

Reliability is the probability of satisfying the target performance. Reliability analysis is applying the reliability estimation methods to the system to achieve better overall performance. Numerous methods for model estimation and evaluation like information criteria can be found in literature. Bootstrap Information Criteria skips the use of analytic expressions for bias correction and instead uses computer generated calculations to estimate the bias correction using bootstrap sampling. Different bootstrap resampling techniques like balanced resampling play a pivotal role in computing parameters of interest and thus affecting the statistical estimation. A unique combination of bootstrap resampling techniques with Bootstrap Information Criteria is achieved to compute reliability and numerical examples are solved.