Activity Number:
|
154
|
Type:
|
Contributed
|
Date/Time:
|
Monday, August 12, 2002 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Quality & Productivity*
|
Abstract - #301550 |
Title:
|
The Effects of Inertia in EWMA and CUSUM Control Charts
|
Author(s):
|
Zachary Stoumbos*+ and Marion Reynolds, Jr. and William Woodall
|
Affiliation(s):
|
Rutgers University and Virginia Polytechnic Institute and State University and Virginia Polytechnic Institute and State University
|
Address:
|
The Janis H. Levin Building, 94 Rockafeller Road, Piscataway, New Jersey, 08854-8054, USA
|
Keywords:
|
EWMA ; CUSUM ; Inertia ; Control Chart ; SPC
|
Abstract:
|
Exponentially weighted moving average (EWMA) control charts are much more effective than traditional Shewhart charts for detecting small sustained changes in process parameters. EWMA charts are comparable to cumulative sum (CUSUM) control charts in terms of most statistical performance measures. One potential disadvantage to EWMA control charts relative to Shewhart and CUSUM control charts is that the EWMA statistic may be in a disadvantageous position when a process parameter changes. For example, the EWMA statistic may be close to the lower control limit when there is a parameter change that should produce a signal at the upper control limit. In this case, the inertia of the EWMA statistic may delay the movement to the upper control limit and prevent fast detection of this parameter change. The effects of this inertia problem on the performance of the EWMA chart are investigated and the performance of EWMA charts is compared to the performance of CUSUM charts. Recommendations are made for the design of EWMA control charts to mitigate the inertia problem.
|