General Ranked Set Sampling With Cost Model

Ranked set sampling (RSS) which is put forward by McIntyre in 1952, is a kind of sampling method that different from the simple random sampling. When the measurement of data bring irreparable damage or costly, and data sorting is relatively easy, ranked set sampling will be the good choice for effective collection of data. But in real life, when we consider to choose which kind of sampling method, it is inevitable that the research funding will be considered, expecting to seek a balance between sampling effect and cost. In such a situation, the ranked set sampling cost problem arises.This paper mainly studies a branch of generalized ranked set sampling-- the problem of cost model under GRSS sampling. In the premise that the sample mean is unbiased estimate of the total mean, the variance of the sample mean is compared with the variance of the simple random sampling with relative accuracy. And then, considering the sampling cost, the model is corrected based on the previous costing model, and the validity of the two samples is evaluated simply; Finally, the scope and validity of the two sampling applications are simply illustrated by simulated numerical values when the general obeys normal distribution and exponential distribution respectively. This paper attempts to apply result of the simulation to practice, so that to achieve a better effect.