Effect of Risk Aversion on Single Sample Attribute Inspection Plans

The impact of a decision maker's risk aversion on the choice of an optimal sampling plan is empirically examined for various batch sizes and prior distributions of the process quality level. These plans are compared to those of the classical linear cost model. The comparisons show that substantial differences in the characteristics of an optimal sampling plan (i.e., probability of acceptance, optimal sample size and expected total cost) occur when risk aversion is considered. A discussion of the optimization characteristics involved in the selection of a sampling plan is also presented. An optimization algorithm which obtains an optimal sampling plan based on the criterion of maximizing expected utility is proposed and evaluated.