Abstract
It is proposed to design single sampling plan by attributes for integrated units based on compound binomial-poisson distribution. It is assumed that an integrated unit is composed of m(≥2) parts and is classified as "defective" if at least one of m parts is defective, The information of the acceptance number of integrated units and that of parts are both used for acceptance criteria. In this paper, the compound distribution of the number of defectives of integrated units and the number of defectives of composed parts is derived. When the strength of sampling plan, say (p_0,α, p_1,β), is given, a sample size and both acceptance number of integrated units and that of parts which satisfy the strength are determined. Furthermore, the sampling tables for sampling inspection plan of integrated unit are given, and sample sizes of the proposed plans are compared to those of plans which are only based on the information of number of defectives of integrated units.
