A comparison of control charts for the average of autocorrelated processes

Control charts are extensively used with the purpose of monitoring some parameters of the process. In general these charts are based on the normality and independence assumptions of the sample observations. However, there are situations where the independence is not valid such as in chemical processes or sampling on-line. In this paper we compared the control charts based on geostatistics and time series methodologies with the well-known charts Shewhart, CUSUM and EWMA, when used to monitor the average of autocorrelated processes. The comparison was performed by using Monte Carlo simulation implemented in the software R for Windows.