Reliability Assessment Using Variance Reduction Techniques Based on Fuzzy System Performance

Classical reliability methods require a crisp definition of system failure according to some performance measure that is of interest to an analyst. Therefore, they do not consider any form of uncertainty in the system failure. This limitation might result in reliability assessments that are not realistic. The uncertainty associated with qualifying the performance of the system as inadequate, i.e., failed, should be considered in reliability assessment. The objective of this contribution is to generalize currently used reliability assessment methods by treating failure as a fuzzy event. Therefore, the reliability of the system can be provided over a nonperformance spectrum. Monte Carlo simulation with variance reduction techniques will be used in the computation of the reliability.