Design of complex DOE

[#1]

I am trying to figure out a way to run an experiment using JMP and I was hoping to get some feedback on how to construct one:

a. I have a formulation of "n" components

b. For some of the "n" components, I would like to investigate the impact of its property. For eg, I have a component "n1" and for this "n1" I have three levels of properties: "n1a, n1b and n1c".

c. Under certain circumstances, the properties of the nth component and its formulation can have an impact so there may be some interactions.

How do I construct such a DOE using JMP? I looked into options such as a nested DOE but couldn't find the right way to build it in JMP. Any help will be appreciated here.

[#2]

Welcome to the community.  I don't think I completely understand your hypothetical, but here are my initial thoughts:

1. Typically when looking to "optimize" n components of a formulation, you would use mixture designs as there is a constraint.

https://www.jmp.com/support/help/en/16.2/?os=mac&source=application&utm_source=helpmenu&utm_medium=a...

2.  What do you mean by "its property"?  Are these nominal (categorical) or continuous levels?  

3.  What circumstances?  I'm confused by your use of the word formulation: " ...the properties of the nth component and its formulation".  I thought you were looking at components that make up a formulation (your bullet 1), but now it appears the component is made of a formation?

4. What are the response variables?  What data type?  Have the measurements systems been evaluated?

[#3]

In addition to @statman , did you look at JMP Help?

Design of Experiments Guide (jmp.com)

And there are many examples under "Help --> Sample Data" regarding DOE. Look e.g. for vinyl, this is an example with mixture components and process factors, with results, probably this would be worth to look at.

Additionally I can recommed the STIPS DOE course: Design of Experiments (DOE) Course | JMP, it's a good introduction, and tells you about JMP's DOE capabilities and which terms are used - terms are important.

And executing the following script will give you an idea how a very simple mixture design in JMP may look like (New Script, paste it and execute, I took this from scripting index).

Names Default To Here( 1 );
d = doe( Mixture Design );
d << Mixture Design Type( Simplex Centroid, 2 );