Improving the Performance of Bi-Variate EWMA Control Charts Using Bayesian Approach

Extended Abstract In many situations, the quality of a process can be characterized by a single continuous random variable, which is usually assumed to follow a normal distribution. However, it is increasingly common for processes to be characterized by several, usually correlated, variables. (Kim and Reynolds 2005) Multivariate control charts are widely used to monitor industrial processes (Mason, et al. 1995). As the objective of performing multivariate statistical process control is to monitor the process over time, in order to detect any unusual events allowing quality and process improvement, it is essential to track the cause of an out-of-control signal. However, as opposed to univariate control charts, the complexity of multivariate control charts and the cross-correlation among variables make it difficult for analysis of assignable causes to the out-of-control signal. This is the basis for extensive research performed in the field of multivariate control chart since the 1940’s, when Hotteling (1947) recognized that the quality of a product might depend on several correlated