摘要
Acceptance sampling is a practical and widely used approach for verifying product quality in modern manufacturing. Conventional two-plan sampling systems (TSS) switch between tightened and normal inspection to balance inspection cost and decision risk, but existing designs adjust only the sample size (TSS-I) or the acceptance benchmark (TSS-II), which limits their flexibility and efficiency. Motivated by this limitation, this study proposes an integrated two-plan sampling system (ITSS) that simultaneously enlarges the sample size and tightens the acceptance benchmark during tightened inspection while remaining compatible with existing TSS designs. The ITSS is built on a one-sided process capability index Cw, enabling efficient lot disposition for largerthe-better and smaller-the-better quality characteristics under high-yield conditions and providing a transparent link between process capability, defect rate, and decision risk. A Markov-chain-based analytical framework is established to derive the operating characteristic (OC) and average sample number (ASN) functions of the ITSS, and an optimization model is formulated to minimize the average sample number at the acceptable quality level subject to producer's and consumer's risk constraints. A cloud-based platform implemented in R and Shiny supports practitioners in configuring design parameters and obtaining optimal ITSS settings without specialized programming effort. Comparative studies demonstrate that the proposed ITSS consistently achieves lower ASN and steeper OC curves than existing Cw-based TSS-I and TSS-II across various design scenarios, and a case study on lithium-ion battery cells illustrates how the ITSS can be deployed in practice to reduce inspection effort while maintaining stringent outgoing quality protection.