Multivariate feed back control: an application in a productive process

The main purpose of this research is to implement a multivariate feedback adjustment proportional to the last deviation from the target, in the set of variables wandering around the target. To apply the controller equation it will be necessary to study the exponentially weighted moving average (EWMA) statistic, in order to determine the behavior of the target disturbances. To determine the forecast values of the variables we will use Seemingly Unrelated Regression (SUR), which is necessary since there is a relationship between the variables and between the errors. In this manner, multivariate feedback adjustment can be reached based on scientific grounds. q 2004 Elsevier Ltd. All rights reserved.