Construction of single sampling plan indexed through average quality level.

Abstract: In certain product categories, there will be defective products in virtually every production lot. It is often true even after the manufacturer has checked each individual product and has revamped the defective ones, since visual inspection is not 100% consistent. Consequently, in many supplier or buyer interactions, the supplier is not expected to deliver defect-free commodities. The buyer needs to control the quality of purchased commodities, since the buyer does not want numerous defects. One of the foremost techniques is ‘Acceptance Quality Limit' (AQL) which has always stood for the quality level explicitly at the worst tolerability limit. The actual definition of AQL is the percentage of defective parts that is routinely accepted by the sampling plan. In this juncture there is a considerable procedure to deal with this problem using Poisson distribution which is often used as a standard probability model for dealing with this kind of single sampling plan. However many data sets in general are not well fitted by a Poisson model, because they consists of more zero counts than are compatible with the Poisson model. Zero inflated models are used to model count data for which proportion of zero counts is greater than expected. For these situations, a zero inflated Poisson (ZIP) model is generally proposed. It is very logical and encouraging that Zero Inflated Poisson Distribution provided better fit as compared to Poisson distribution in the data sets because of more number of zeros. In this paper the procedure for the construction of Single Sampling Plan (SSP) indexed through Average Quality Level (AQL) using Zero Inflated Poisson (ZIP) distribution as the base line distribution is presented and a table is also presented using Excel packages for the easy selection of the plans.

Back to List of publications / Home