A mathematical model for multiattributes acceptance sampling systems is presented. Through this model the expected total cost of quality control per lot can be determined under the assumption that rejected lots are scrapped. The lot proportion defective for each attribut, e type is treated as an independent random variable. Any appropriate distribution may be used to describe the random nature of proportion defective for a given attribute type. Optimization of the model is achieved through a well known search technique and is illustrated through an example problem. Sensitivity analysis of the optimum to changes in the assumed prior distributions is presented. 1972 Taylor & Francis Group, LLC.
