Serially dependent process double sampling plans

The design of acceptance sampling plans in case of serially dependent stochastic process is generally very complex. In this paper an efficient design procedure and estimation, of the double sampling attribute plans are developed for the serially dependent process models. If any dependent process model can be simulated, the best sampling plans can be selected efficiently by using the special properties of the probability structure. The number of actual evaluations to find the best sampling plans can be reduced remarkably. The experimental study reveals that only small portion of the total exhaustive enumeration is needed. A Markov chain model is given as a numerical example.