Sankhya: The Indian Journal of Statistics
1999, Volume 61, Series B, Pt. 2, pp. 351-365
A STATISTICAL MODEL FOR THE DESIGN OF MULTIATTRIBUTE CONTROL CHARTS
JOEL K. JOLAYEMI, University of Zululand, KwaDlangezwa
ATANU SENGUPTA, Indian Statistical Institute, Calcutta
SUMMARY. We develop a statistical model for the design of multivariate control charts (MACCS). A formula for determining quick solutions to the model to obtain the charts' design parameters is also developed. Furthermore, we develop a procedure for identifying attributes that are responsible for alarm signals given by the charts. Numerical examples are given and OC curves are drawn for some of the designs produced in the examples to illustrate the model and to compare the performances of MACC with ordinary np-charts. The examples show that the new MACC technique is more cost-effective and more efficient than the np-chart technique. The former produces smaller sample sizes than those produced by the latter. The MACC is also shown to be, oftentimes, more powerful than the np-chart.
AMS (1991) subject classification.62N10.
Key words and phrases. Multiple attributes, convolution, independent binomial variables, control charts, acceptance number, operating characteristic curves.
Full paper (PDF)