Consider the nonlinear model y = f(x1; x2) = 1e2x1 − 3e4x2 .

(a) Linearize the model by taking the partial derivatives @f

@i

; i = 1; 4.

(b) Construct a grid of candidate points in the experimental region range 1 B x1 B 7; 1 B x2 B 7.

(c) Use the OptFederov function to construct a D-optimal nonlinear design for this model using the initial estimates 1 = 10:0, 2 = −0:5, 3 = 5:0, and 4 = −2:0.