Optimal Bayesian reliability assurance test plan for degraded products based on Wiener process

Reliability assurance tests have been widely adopted to determine whether products meet a specified reliability level under a prescribed confidence level. However, for highly reliable products, insufficient failure data are often observed during assurance testing. To address this issue, this paper proposes an optimal reliability assurance test plan for degraded products using a Wiener process, accounting for unit heterogeneity and measurement error. Single and double sampling plans are constructed, with posterior producer’s and consumer’s risks defined in a Bayesian framework. Two optimization models are proposed: one minimizes the sample size or average sample number under posterior risk constraints, and the other minimizes the expected cost associated with decision uncertainty and subsequent reliability assurance activities. In addition, a dynamic strategy for adjusting the nonconforming threshold over time is investigated. Numerical analyses are conducted to compare the single and double sampling plans and to examine how reliability requirements, risk constraints, and test conditions affect the optimal designs. A case study further demonstrates the effectiveness and applicability of the proposed method.