A new method for phase II monitoring of multivariate simple linear profiles

TL;DR

This paper introduces an EWMA chart for phase II monitoring of multivariate simple linear profiles, effectively detecting moderate and large shifts in correlated response variables, improving quality control processes.

Contribution

It proposes a novel EWMA-based monitoring scheme specifically designed for multivariate linear profiles with correlated responses, enhancing detection capabilities over previous methods.

Findings

Superior detection of moderate and large shifts

Effective in monitoring correlated response variables

Outperforms previous methods in key scenarios

Abstract

A scope in quality control, which has recently received a great deal of attention is profile that characterizes the quality of a product or process by a relationship between two or more variables. In this paper, we propose an EWMA chart for phase II monitoring of multivariate simple linear profile in which several correlated response variables have linear relationships with one explanatory variable. The statistical performance of this scheme is evaluated in terms of out-of-control average run length. Although it seldom signals for small shifts, it is superior to previous works in detecting moderate and big shifts.