Consider the experimental design problem in which there are three factors: (1) impeller speed (in rpm) with associated variable , such that the lower and upper levels are 250 and 400, respectively, (2) amount of binding solution (in %) with associated variable , such that the lower and upper levels are 60 and 70, respectively, (3) pelletization time (in min) with associated variable , such that the lower and upper levels are 5 and 20, respectively.
Pelletization experiments using Doehlert design scheme determine one parameter: the mean pellet size (in mm).
This Demonstration computes the quadratic model by multilinear regression; the response versus , , and is given by:
You can set the value of one of the variables (say ) and the Demonstration will display the contour plot of the response versus the other two variables (and ).
G. A. Lewis, D. Mathieu, and R. Phan–Than–Luu, Pharmaceutical Experimental Design, Drugs and the Pharmaceutical Sciences, Vol 92, New York: Marcel Dekker, 1999.