摘要
•Propose an adaptive inspection scheme for high-yield quality verification.•Enhances cost-efficiency via historical lot records and repeated sampling control.•Develops a nonlinear optimization model to determine optimal plan parameters.•Demonstrates superior performance compared with existing sampling plans.•Provides an interactive online solver to support practical implementation.
Acceptance sampling plans (ASPs) effectively verify product quality in practice. With the advancement of sampling strategies, ASPs have evolved from basic single sampling plans to more cost-effective approaches, such as multiple dependent state sampling plans, which review historical lot quality records, and repetitive group sampling plans (RGSPs), which employ repeated sampling. More recently, scholars have combined these approaches to propose a modified RGSP (MRGSP) to further enhance cost-effectiveness. However, the MRGSP still exhibits notable drawbacks: (i) its cost-effectiveness decreases as more historical lot quality records are reviewed, and (ii) it lacks a limit on the number of repeated samplings, potentially leading to excessive or even infinite resampling. To address these issues, we propose an adaptive inspection scheme (AIS) that integrates a flexible and adaptive mechanism for reviewing lot quality history and managing repeated sampling. Extensive analyses and comparisons demonstrate that the AIS effectively overcomes the deficiencies of the MRGSP and significantly improves cost-effectiveness compared with existing plans, while maintaining reliable discriminatory power in lot quality decisions. Furthermore, because solving the optimization model of the AIS is challenging due to its adaptive sampling mechanism, we develop an online solver to assist practitioners in efficiently obtaining optimal design parameters and facilitating practical implementation. Finally, a case study illustrates the applicability of the proposed AIS.