Year
1987
Abstract
STRACT Traditionally, sampling plans have been based either on randomized sampling or have relied on best engineering judgment. Here, a new strategy is proposed using differences in the risks associated with different units of the sampling set (e.g., a waste barrel, a unit surface at a work or contamination site) to influence the frequency of sampling. This results in a sampling plan which, unlike conventional random sampling, can form the basis for tests involving risk considerations. Sampling rates are higher for units which pose a higher risk, while maintaining a sufficient rate for units with a much smaller risk. This strategy is presented here in a general form and examples are given in the area of transport and disposal of nuclear waste.