ASN-Minimax double sampling plans by variables for two-sided specification limits when σ is unknown (1103.4801v4)

Abstract: ASN-Minimax double sampling plans by variables for a normally distributed quality characteristic with unknown standard deviation and two-sided specification limits are introduced. These plans base on the essentially Maximum-Likelihood (ML) estimator p* and the Minimum Variance Unbiased (MVU) estimator ^p of the fraction defective p. The operation characteristic (OC) of the ASN-Minimax double sampling plans is determined by using the independent random variables p*_1, p*_2 and ^p_1, ^p_2, which relate to the first and second samples, respectively. The maximum of the average sample number (ASN) of these plans is shown to be considerably smaller than the sample size of the corresponding single sampling plans.