From: Stan Hilliard
Date: 31 Oct 2002
Time: 10:12:42
1) For Ac=5, you accept the lot if the sample contains 5 or fewer defectives.
2) There is no reason to limit AQL and RQL to be whole numbers.
3) For the choice of alpha and beta risks (your alpha=1-.95 & beta=.10) I prefer the practice of alpha=beta=0.05. Imagine that you are the producer and have to defend your plan in a product liability lawsuit. The judge will ask you why you chose a consumers risk that is twice as large as your own producer's risk. Try to come up with a non-selfish answer to that!
4) The values of n=100, Ac=5 are the decision rule of the sampling plan for deciding whether to accept or to reject a lot. The choose/design a decision rule that is valid for your application requires that you examine its decision making PROPERTIES. Those properties of a decision rule are completely described by the rule's operating characteristic curve.
samplingplans.com/modern3.htm#EVALUATE