An Economic Choice of Sampling Inspection Plans Under Varying Process Quality

The costs of orthodox sampling plans, namely single‐sampling, double‐sampling and multiple‐sampling plans together with some plans that utilize neighbouring batch information, i.e. the dependent double‐ and deferred double‐sampling plans of Baker and Brobst, chain‐deferred sampling and multiple‐chain‐deferred sampling plans of Osanaiye, are compared. This is done under a linear trend in process quality, an autoregressive process and an inert process. Weibull's cost model was adopted in obtaining the cost per accepted effective item and the various parameters used were obtained by simulation and sentencing the simulated lots accordingly. The effects of varying the cost parameters on the overall cost per accepted effective item are examined by plotting each against a wide range of parameter values. The multiple‐chain‐deferred and the multiple‐sampling plans are generally cheaper than all the other plans in the comparison, but where they are not used because they are more complicated the chain‐deferred sampling plan is generally the best.