cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
  • Register to attend Discovery Summit 2025 Online: Early Users Edition, Sept. 24-25.
  • New JMP features coming to desktops everywhere this September. Sign up to learn more at jmp.com/launch.
Choose Language Hide Translation Bar
Here4DOE
Level I

I-optimal custom DOE RSM design with continuous and categorical factors - centre points not balanced?

Hi, I want to create an I-optimal custom DOE with 5 continuous factors and one categorical factor (with 2 levels). I also want to have 4 centre points which I specify on set-up. I also choose RSM for my model as I know there is curvature. When JMP creates the design, the centre points are not equally divided across the two categorical levels - one category gets 4 centre points and the other gets 2 centre points. I would manually alter them to be balanced, however, this would result in L1 having two more runs throughout the experiment than L2. Is there anything that I can do to ensure the centre points are balanced across these two levels?

DOE Set-UpDOE Set-UpDOE TableDOE Table

1 REPLY 1
Victor_G
Super User

Re: I-optimal custom DOE RSM design with continuous and categorical factors - centre points not balanced?


Hi @Here4DOE,

 

The "quick fix" solution could be to create the design without the centre points (so that levels of the categorical factors are balanced), and manually add the 6 centre points (3 for L1, 3 for L2). You could then randomize your design table by using the Create a Subset Data Table platform with sampling rate of 1, to keep all rows from your design but randomize the order.


Note that if you really need to enforce a balance between your 2 levels of categorical factor, you need to have an even number of total runs (like 32 or 34 in your example). Else JMP will try to find the best trade-off/balance between the repartition of centre points and the "other" points so that all runs are distributed in a "balanced" way across all levels.


Hope this answer can help you,

Victor GUILLER

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

Recommended Articles