Accelerated acceptance sampling plan for degraded products based on inverse Gaussian process considering the acceleration factor uncertainty

Abstract

Acceptance sampling plan based on accelerated degradation tests, denoted as ADASP, is widely employed to verify the reliability of degraded products, including risk analysis, sampling plan design, and decision-making criterion determination. Most existing research focuses on identifying the optimal accelerated degradation test to improve the efficiency of ADASP, but often overlooks the differing objectives of producers and consumers regarding the acceptance index, which may result in the test plan failing to effectively meet their needs. To address this, based on the inverse Gaussian process, we propose a comprehensive accelerated degradation sampling plan by optimizing the parameter estimation accuracy while protecting their interests. Beyond the product quality risk, the acceleration factor (AF) uncertainty introduces additional risk to ADASP. Current studies primarily tackle AF uncertainty using probability distribution or interval. However, obtaining AF’s distribution is challenging, especially with limited prior knowledge and complex models. Therefore, apply the generalized pivotal quantity to derive a confidence interval of AF, and effectively manage this uncertainty-driven risk using significance level. Subsequently, a detailed decision-making criterion is derived by solving the risk constraint equations for both parties. Finally, simulation studies and case applications on springs are conducted to demonstrate the effectiveness of the proposed method.