Title: An effective chart to monitor process averages for serial correlation using ANN approach
Authors: D.R. Prajapati; Sukhraj Singh
Addresses: Department of Mechanical Engineering, PEC University of Technology, (formerly Punjab Engineering College), Chandigarh-160012, India ' Department of Mechanical Engineering, PEC University of Technology, (formerly Punjab Engineering College), Chandigarh-160012, India
Abstract: Control charts, one of the tools of quality control, used to detect normal and unusual variation/s in a process. The performance of the chart is measured in terms of the average run length (ARL), which is the average number of samples before getting an out-of-control signal. In this paper, the ARLs at various sets of parameters of the X chart are computed by simulation, using MATLAB. An attempt has been made to counter the effect of autocorrelation by designing the X chart, using sum of chi-squares theory. Various optimal schemes of modified X chart for sample sizes (n) of 2 and 4 are proposed at different levels of correlation (Φ). Moreover, these optimal schemes are also validated and compared with the ARLs obtained by artificial neural networks (ANNs). It is concluded that the modified X chart offers more robustness for autocorrelation.
Keywords: average run length; ARL; modified x-bar charts; sample size (n); artificial neural networks; ANNs; level of correlation; control charts; quality control; simulation; autocorrelation.
DOI: 10.1504/IJISE.2015.070873
International Journal of Industrial and Systems Engineering, 2015 Vol.21 No.1, pp.31 - 49
Received: 02 Sep 2013
Accepted: 19 Jan 2014
Published online: 31 Jul 2015 *