Abstract: This paper investigates the effect of the allocation of
sampling resources to the efficiency of control chart. In
spite of statistical properties and quality loss, the initial performance, the
steady state performance and the optimization results show that under the
condition of the fixed sampling rate, for the large process mean shifts or large
expectation of random shifts, we should take the samples of small size
frequently; for the small shifts or small expectation of random shifts, we can
take the samples of large size.
Keywords and phrases: control charts, average run length, statistical process control, control limits, Taguchi loss function.