Page 185 - Special Topic Session (STS) - Volume 3

STS539 Muhammad Abid et al.

A new nonparametric homogeneously weighted
moving average control chart for monitoring the
process location
1
3
2
Muhammad Abid , Hafiz Zafar Nazir , Muhammad Riaz
1 Government College University, Faisalabad, Pakistan
2 University of Sargodha, Pakistan
3 King Fahad University of Petroleum and Minerals, Saudi Arabia

Abstract
The main advantage of homogenously weighted moving average (HWMA)
control chart in comparison to the exponentially weighted moving average
(EWMA) control chart is that the plotting statistic of HWMA chart assigns
specific weight to the current observations and the remaining weights are
equally distributed between the previous observations. This study suggests a
new non-parametric HWMA chart using an arcsine transformation for
monitoring the process target. To compute the average run lengths profile a
Monte Carlo simulations are used. The dominance of the proposed chart is
constructed against its competitors such as nonparametric EWMA sign, EWMA
arcsine, CUSUM sign and mixed EWMA-CUSUM arcsine charts. The study
found that the proposed chart performs efficiently for detecting small and as
well as larger shifts in process target.

1. Introduction
Statistical process monitoring (SPM) consists of several tools which are
used to monitor, control and improve the quality of a product. From these
SPM tools control chart is an important tool to simplify process control.
Control charts are categorized into memoryless and memory control charts.
Memoryless control charts only utilize the up-to-date information in the
statistic but the memory control charts develop on the basis of past
information along with up-to-date information. Shewhart control chart
1
belongs to the category of memoryless charts and performs efficiently to
detect large shifts in process parameters. The cumulative sum (CUSUM) by
3
2
Page and the exponentially weighted moving average (EWMA) by Roberts
charts belongs to the type of memory charts and useful to detect smaller shifts
in the process location or/and variation.
In the literature, several kinds of modifications of the charting strategies
4
have been proposed under various setups. Lucas and Crosier and Lucas and
5
Saccucci applied the fast initial response (FIR) on CUSUM and EWMA charts,
6
7
respectively. Abbas et al. and Zaman et al. introduced mixed EWMA-CUSUM
and mixed CUSUM-EWMA charts, respectively. These charts showed quicker
small shifts detection ability against the EWMA and CUSUM charts. Abid et

174 |I S I W S C 2 0 1 9