Usually, traditional control chart methodology
is based on the standard assumption that random observations are statistically
independent and uniformly distributed. However, for the random data of interest
in practical applications the observations are usually serially correlation.
In many practical processes such as in chemical processes, the random variables
are always serially-correlated. In the case of daily flow of a river, wind
speeds, or the amount of dissolved oxygen in a river, most process data are
autocorrelated.