DoE: disallowed combinations filter gives unexpected result

matthias_bruchh

Hi there,

 

I'm making a testplan for a multilaboratory study on mechanical properties of steel. One of the parameters of interest is "surface roughness" which can have three values (-1,0,1). However for technical reasons the highes surface roughness can only be applied to samples with a minimum diameter of 6 mm. To take that into account, I use the disallowed  combinations filter :

Diameter >= 0 & Diameter <= 5.625 & (Surface roughness >= 0.4306 & Surface roughness <= 1)

This should preclude any tests where the specimen diameter is below 5.625 and the surface roughness above 0.4306. (The decimal number come from using the sliders. There are only a limited number of diameters and 3 surface roughness values, so the excat numbers don't matter.)

 

Nevertheless the test plan comes up with a test (#32 in the attached file) with surface roughness 1 for a diamter of 4.5. (Which according to my understanding should be excluded.)

 

Do I misunderstand something?

 

I attach the Jmp script for reference. (I normally don't use scripting, I just saved the script for this post.)

 

Session Script.jsl

Ryan_Lekivetz

Apologies that you came across this issue. You are correct in your understanding - unfortunately, this is an issue that can sometimes occur with covariates and disallowed combinations. For this particular example, I would manually change surface roughness to a valid value (I think either choice shouldn't make a big difference, but if you're concerned, you can try Evaluate Design or Compare Designs to check).

 

I've made a note to get this addressed for a future release.

Cheers,

Ryan