DISCRETE ECONOMIC MULTIATTRIBUTE ACCEPTANCE SAMPLING.

A mathematical model for multiattribute acceptance sampling is developed. The number of defectives of each attribute in the lot is considered an independent random variable and is properly described by a discrete prior mass function. The concept of distributional reproducibility to hypergeometric sampling is invoked to simplify the expressions normally associated with a discrete model; yet, the model remains exact and may be efficiently optimized. The acceptance sampling situation modeled is identical to one previously published which applied continuous distributions to approximate the lot fraction defective. An extensive example problem is presented in which the two modeling approaches are compared using sensitivity measures.