Studies aiming at the optimization of experimental variables are generally carried out by procedures that evaluate the effect of just one variable on time. However, this approach does not allow the establishment of true optimal values, mainly due to close connections that may exist among the studied variables. In this paper some practical applications based upon factorial design systems are used to demonstrate the convenience of multivariate procedures. The three examples described here entail the optimization of degradation processes related to the experimental design as well as interpretation of results.
Experimental optimization; multivariate analysis; factorial design
