Abstract
A mathematical programming model is developed for the design of sampling plans under varying inspection errors. The model minimizes average lot inspection cost subject to a given outgoing lot quality. An efficient search method is developed and is compared to complete enumeration based on six varying error models. The method always finds a global optimum and its efficiency increases on the maximum sample size. For a set of test problems, the method has examined at most 3.63% of all sampling plans to find a global optimum. This study verifies that the assumption of constant error or no error must be reconsidered. Also, the fact that the sampling plan must be adjusted depending on the inspection error models is justified.
Handled by the Department of Quality and Reliability Engineering.
Notes
Handled by the Department of Quality and Reliability Engineering.