摘要
Acceptance sampling has been used as a method of determining if submitted product lots meet required quality standards. Recently, several sampling plans with different strategies have been developed to tackle various types of situations and requirements, and to increase their performance and efficiency. In particular, repetitive group sampling plan (RGSP) has been proposed and proven to be an efficient sampling scheme. However, it has its drawbacks that occur when the sampling is repeated until the lot can be successfully determined as either accepted or rejected; in theory, the number of sampling times might be infinite. Therefore, this paper proposes a modified version of variables RGSP (VRGSP) with an adjustable mechanism for lot sentencing (Adj-VRGSP), i.e., by setting a maximum allowable sampling time. The operating characteristic (OC) and average sample number (ASN) functions of the proposed plan are derived along with the step-by-step operating procedure. The determination of plan parameters is formulated as an optimization model with nonlinear constraints related to quality and risk requirements. Additionally, the performance of the proposed plan is examined, discussed, and compared with that of the conventional sampling plans. An application is also presented as an example illustrating the effectiveness of the proposed plan.