I have a challenging profiler question.

[#1]

I have mixture data in which one of the ingredients has three categories- "Type A", "Type B", and  "None".  "None" wasn't part of the original DOE but was manually augmented.  Is it possible to constrain the profiler settings from any number >0 when "None" is selected?  Attached is an example file.

[#2 · Accepted solution]

You can fix a value manually in the profiler.  One thing I did notice is that since you set up your design using JMP there are column properties on your Xs that constrain the acceptable ranges.  So for X1 the low limit is 0.2.  Since you added runs manually you will need to adjust these limits to include the new range.  Once you fix these ranges your profiler will go down to zero on the X1.

[#3]

But after I adjust these limits, there is no constraint on values greater than zero when "None" is selected.  Once "None" is selected, X1 should be constrained to 0.

[#4]

Thank you for your help. 

[#5]

The profiler only recognizes linear constraints involving continuous factors.  You can "force" the profiler to do what you'd like with Karen's suggestion.  As she mentions you need to change the lower limit of X1 to go to 0.

Then you'll need to lock X1 at 0.  The line in the X1 profile will turn solid red once it is locked at a value.

Note that if you want to change X1 Type to "A" or "B", you will probably need to unlock X1 from zero.  Currently, there is no way to have the profiler do this automatically.  I hope this helps.