Optimal operational selection and sequencing of systems with mixtures of one-shot units

Abstract

One-shot units are produced and stored in batches at arbitrary times. Non-destructive tests (NDTs) are conducted randomly and repeatedly at arbitrary times during the one-shot units' entire life horizon to assess its reliability metrics. Some units (with nonhomogeneous characteristics) are selected from storage and launched or operationally used at arbitrary times as needed. Referring to the launched units as a system, we optimize the system's operational use by determining the selected units' characteristics (number of units selected from each batch) and its launching sequence. The system reliability metrics such as the expected number of successfully launched units, the average and variance of the probability that the system achieves a successful operation, and the expected time of the system's kthfailure are considered in the optimum selection procedure. Considering the units' inhomogeneity, we provide the confidence bounds of the probability that the system achieves a successful operation based on the units' optimal selection and launching sequence. A simulation model is developed to validate the proposed optimal sequence.