Utilizing contour map to set up the constraints for Augment Design

[#1]

Hi community,

I am working on a formulation work. I got initial 3 by 3 Design as below with replicates.

Then I got decet predictive model.

Below is the controur plot with raw data vs plot predicted by the model. They look alike except the region where there is no data points (white space),

Now I want to optimize this formulation by further looking into the high performing formulation (both upper left and middle).

Upper left can be easily zoomed in through linear constraints. However I am not sure what I should do to look into "hot red" region in the middle. Does anybody know easy way to apply constraints on this region?

Thanks

[#2]

Quick question:  If this is a formulation, shouldn't you be running Mixture Designs?

https://www.jmp.com/support/help/en/17.2/?os=mac&source=application#page/jmp/mixture-designs.shtml

[#3]

Hi @statman 

That is a reasonable question to ask. Based on our prior knowledge in this formulation, we know that the ratio of two excipients (factor1) and the total concentration (factor2) are the two critical factors that affect the readout.

I admit, though, the composition (or relative amount) will change as well by changing those two parameters.