Fractional Failures Sampling-An Alternative to Zero Failures Sampling Plans

Zero failures sampling plans have been widely established in quality management and control. Their performance is better than sampling plans that allow failures. This is reflected in the operating characteristic (OC) curve. The OC-curve gives the likelihood of acceptance of a random sample as a function of the proportion of defective devices in a population. In this paper, we add information to the sampling plan. If a failure occurs in the sample and a countermeasure is implemented, we consider the effectiveness of this countermeasure as well. In reliability engineering, this concept is known as fractional failures. Taking this additional information into account and increasing the size of the random sample leads to sampling plans that have at least the same performance as zero failures sampling plans. The consumer's risk is controlled as well as the producer's risk. The results are very easy to handle and very transparent in the communication between producers and their customers. This method gives an alternative way how zero failures sampling plans can be realized and assessed.