Define Factor Constraints; Do certain factor types prevent use?

Question

Hi,

I'd like to pick all of your brains on something.

So, I'm currently producing a buffer screen where I am trying to keep the design space within a specified Osmolality limit (250-350 mOsmo/kg).

I'm adding my factors, 1 catagorical, 1 Discrete Numeric and 7 continuous as normal, however when I try to 'Specify Linear Constraints' I cannot build a design.

One of two things happens; either I don't see the linear constraints options when I try to to add them and then i'm not able to progress the design, OR I CAN see the constraints (in which case I add a <350 and >250 line for each), but then it also fails when I try to make the design.

I get one of the following error messages :

"invalid disallowed combinations script. See log for details" or

"The selected terms in the Model outline are linearly dependent on the previous terms. Please press the remove term button and try agian."

If I try the design using everything as a continuous factor there's no issue (though obviously that doesn't suit my design).

Now, you'll be relieved to hear that I managed to build the design by using the 'Disallowed Combinations Script' after writing code to keep the design within my limits, but I am now curious as to why I couldn't use the more straighforward approach. Is there a conflict with the catagorical or discrete numeric factors that prevents my use of the simpler 'specifiy linear constraints' option, or was I doing something else (and possibly exciting) wrong?

Thanks

Jon

Mark_Bailey · Accepted Answer

You are not doing anything wrong. It is simply a matter of what the user interface allows you to do. There are several options. The linear constraints, as you know, is straight-forward but it only with continuous or mixture factors. It is a convenience for a simple, special case. Your case is more general with other types of factors for which the linear constraints was not designed. The use of Boolean JSL expressions is the most general solution using the disallowed combinations approach.

I guess you will have to look elsewhere for more excitement.