摘要
Acceptance sampling plans are preferably used when sample testing becomes destructive or costly to make reliable lot determination from a reasonable size of samples. Previous decades have seen various sampling strategies developed for specific purposes and the evolution of sampling methods continue to ensure that lot sentencing becomes more efficient and economic. In this study, we proposed a modified variables repetitive group sampling (VRGS) plan integrated with a critical-value-adjusted switching mechanism. Such a design provides an adaptive assessment of the varied submitted quality level. The operating characteristic function is derived from the sampling distribution of process yield index S pk , which possesses a one-to-one transformation to process yield. By solving the developed nonlinear minimization model, the plan parameters of the proposed method are satisfied with tolerable sampling risks and desirable quality levels simultaneously. Furthermore, by comparing the average sample number (ASN) reduction and discriminatory power with other approaches, the proposed method outperforms conventional methods. Numerical illustrations are provided to show its practicability in real applications.