cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Try the Materials Informatics Toolkit, which is designed to easily handle SMILES data. This and other helpful add-ins are available in the JMP® Marketplace
Choose Language Hide Translation Bar
DOE_USER
Level III

Augment DOE

Hello,

 

I am trying to set-up a DOE with 5 variables. Some of the combinations are not allowed, and for a regular DOE, I can remove them from consideration using linear constraints. However, if I start with some data, and then use Augment DOE option, it removes my constraints. Is that a feature of JMP, or is there a reason for it? Why can't I have all the features available in custom DOE in Augment DOE?

 

Thank you,
Sweta

1 ACCEPTED SOLUTION

Accepted Solutions
Victor_G
Super User

Re: Augment DOE

Hi @DOE_USER,

 

Indeed, if you try to use blocking with constraints when augmenting the design, you'll receive a warning message :

Victor_G_0-1730795364196.png

You could generate the design you want in two steps :

  1. Generate the new runs without blocks first with the Augment Design platform,
  2. Use these new runs as Candidate set in the Custom Design platform (without changing the assumed model) to create the blocks, and add a blocking factor in the "Factors" panel. Check the option "Include all selected covariate rows in the design" if you want to make sure that all previously generated runs in step 1 will be present in the final design with blocks.

Make sure you have enough runs when first augmenting the design so that the block creation does not create new runs.

 

Hope this answer might help you,

Victor GUILLER

"It is not unusual for a well-designed experiment to analyze itself" (Box, Hunter and Hunter)

View solution in original post

5 REPLIES 5
Victor_G
Super User

Re: Augment DOE

Hi @DOE_USER,

 

Could you provide an example with an anonymized dataset ?
I tried to reproduce a Custom design with 5 continuous factors and 2 linear constraints :

Victor_G_1-1730757204106.png

Generating this design creates a datatable with a "Constraint" script : 

{1 * :X1 + 1 * :X5 <= 5, 1 * :X2 + 1 * :X3 >= 5};

When augmenting this Custom Design with the platform Augment Design, the "Constraint" script is taken into consideration :

Victor_G_0-1730757166886.png

 

So I suppose that if you have started with some data, the "Constraint" script was not present in your datatable, so JMP is not able to use this linear constraint and take it into consideration for the augmentation.

You could try to set this script yourself to "force" JMP to respect your experimental space constraints.

 

I hope this answer will help you,

Victor GUILLER

"It is not unusual for a well-designed experiment to analyze itself" (Box, Hunter and Hunter)
DOE_USER
Level III

Re: Augment DOE

Hello,

 

I realized my constraints are removed if I try to use blocking! (Use block for new runs). If I don't do that, you are right, the constraints are present.

 

Thank you for your response!

Sweta

DOE_USER
Level III

Re: Augment DOE

I also have categorical variables in my DOE (as well as continuous ones). 

Victor_G
Super User

Re: Augment DOE

Hi @DOE_USER,

 

Indeed, if you try to use blocking with constraints when augmenting the design, you'll receive a warning message :

Victor_G_0-1730795364196.png

You could generate the design you want in two steps :

  1. Generate the new runs without blocks first with the Augment Design platform,
  2. Use these new runs as Candidate set in the Custom Design platform (without changing the assumed model) to create the blocks, and add a blocking factor in the "Factors" panel. Check the option "Include all selected covariate rows in the design" if you want to make sure that all previously generated runs in step 1 will be present in the final design with blocks.

Make sure you have enough runs when first augmenting the design so that the block creation does not create new runs.

 

Hope this answer might help you,

Victor GUILLER

"It is not unusual for a well-designed experiment to analyze itself" (Box, Hunter and Hunter)
DOE_USER
Level III

Re: Augment DOE

Thank you Victor!!