An optimal design of CUSUM control charts for binomial counts
作者:
F. F. Gan,
期刊:
Journal of Applied Statistics
(Taylor Available online 1993)
卷期:
Volume 20,
issue 4
页码: 445-460
ISSN:0266-4763
年代: 1993
DOI:10.1080/02664769300000045
出版商: Carfax Publishing Company
数据来源: Taylor
摘要:
The Shewhart p-chart or np-chart is commonly used for monitoring the counts of non-conforming items which are usually well modelled by a binomial distribution with parameters n and p where n is the number of items inspected each time and p is the process fraction of non-conforming items produced. It is well known that the Shewhart chart is not sensitive to small shifts in p. The cumulative sum (CUSUM) chart is a far more powerful charting procedure for detecting small shifts in p and only marginally less powerful in detecting large shifts in p. The choice of chart parameters of a Shewhart chart is well documented in the quality control literature. On the other hand, very little has been done for the more powerful CUSUM chart, possibly due to the fact that the run length distribution of a CUSUM chart is much harder to compute. An optimal design strategy is given here which allows the chart parameters of an optimal CUSUM chart to be determined easily. Optimal choice of n and the relationship between the CUSUM chart and the sequential probability ratio test are also investigated.
点击下载:
PDF (872KB)
返 回