On Designing of Mixed Double Exponentially Weighted Moving Average –Tukey Control Chart Based on Range

Authors

  • Saowanit Sukparungsee Department of Applied Statistics, Faculty of Applied Science, King Mongkut's University of Technology, North Bangkok
  • Supicha Nukvira Department of Applied Statistics, Faculty of Applied Science, King Mongkut’s University of Technology North Bangkok
  • Yupaporn Areepong Department of Applied Statistics, Faculty of Applied Science, King Mongkut’s University of Technology North Bangkok

Keywords:

Nonparametric Control Chart, Dispersion, Average run length, Monte Carlo simulation method

Abstract

This objective of this research is to present a mixed double exponentially weighted moving average-Tukey control chart (DEWMA-Tukey) using a range to detect variation in process variability, when the process is symmetric and asymmetric. And to compare performance in detecting variation among Tukey's Control Chart (Tukey), Double Exponentially Weighted Moving Average charts (DEWMA),  Exponentially Weighted Moving Average-Tukey Control Chart (EWMA-Tukey) and the Double-Exponentially Weighted Moving Average-Tukey Control Chart (DEWMA-Tukey). Criteria used to determine performance is the out of control average run length  when in control average run length is set at 370 and 500. From Monte Carlo simulation found that when the value and the sub sample size (n) are equal to 1, the DEWMA-Tukey control chart is more effective in detecting changes in the distribution value than the EWMA-Tukey control charts, DEWMA control charts, and Tukey control charts when the process is symmetric distribution. And for asymmetric distribution and the sub sample size (n) are equal to 5, DEWMA-Tukey control charts are better than other control charts when the change size is small  and when the sub-sample size (n) is 1. EWMA-Tukey control charts are better than other control charts.

Downloads

Published

2019-06-19

Issue

Section

Research Articles