DOE - Special Purpose - Space Filling Design & Integers

[#1]

Hi JMP community,

I am facing a problem with JMP13. I want to create a DOE - Special Purpose - Space Filling Design in order to generate sets of inputs to be used in a another software model. The problem is that I have 2 variables L, H and they have to:

- Be integers (both)

- There sum should be lower or equal than a certain fixed value lets say X

I dont see a way to indicate that these two variables have to be integers in the pop up window. I found an approach to include those variables as "categorical factors" but if you do that then you dont get to constrain them in the "Define factor constrains", they do not appear as options.

Any suggestions how can we solve these kind of issues in JMP?

Thanks

[#2]

You could consider creating the design with the sum constraint. After the design is created, you might then round the values to the nearest integer. However, after rounding, you would need to check to be sure the constraint still holds for all rows.

[#3]

Hi Susan and thank you for your reply.

Please understand that I am totally new in JMP its the first time that I am using it, so please provide also some guidance, photos etc on your replies if you please.

As I am using categorical variables, I can not add a sum constrain.... I dont fully understand your recommendation.

Can I follow a custom design or a different approach from inside the software without using 3rd party tools like excel for example?

Thanks

[#4]

I would just design the experiment as if the factors were continuous and then convert the factor levels to integers. I think that's the easiest approach.

[#5]

I recommend that you read the DOE Guide.  It very clearly details with examples the many ways to use DOE in JMP.

     Help==>Books==>Design of Experiements Guide

[#6 · Accepted solution]

I am sorry my original answer was not clear. My suggestion was to use the Space Filling Design platform, setting your factors as continuous. This would allow you to add the sum constraint for those continuous factors. You can then take the data table created by the platform and round all values to the nearest integer using formulas. You will need to check the sums of the factors again to be sure the rounding has not violated the constraint, making adjustments as needed.

I hope this helps to clarify.