Abstract:
|
Shewhart control charts for variables data are mainly based on the assumption that the observations are iid and follow a normal distribution. However, in practice, these assumptions may be violated, and this may produce chart performance that may not be expected. This paper explores the robustness of a number of popular Shewhart charts, including the X-bar chart and the S chart, when some of the underlying assumptions are violated. The performance of these control charts will be examined for data simulated from several distributions of different shapes.
|