Can continuous numerical variables with poor adjustability in practice be set as discrete numerical variables

[#1]

Can continuous numerical variables with poor adjustability in practice be set as discrete numerical variables? Because when using the optimal experimental design, the factors given in the design table are not integers, but some parameters can only be controlled as integer values in the experimental process

[#2]

Welcome to the community.  I'm not sure I understand your question and I don't understand the specific situation you are dealing with. Do you mean you have difficulty controlling the variable at a given level for the duration of an experiment?  Or are you questioning the coding of levels for the data table?  I don't understand the integer question at all...sorry.  Perhaps someone else will.

The assumption in experimentation is the factors that are being manipulated can be set at specific levels for the duration of the experiment (or treatment).  Now we all know there could be some variation around the setting and should plan the difference between levels accordingly (e.g., bold level setting).  If the variable can't be controlled, but can be measured, perhaps using regression would be helpful.

[#3]

Yes, you may use Discrete Numeric factors in such a case. Just be sure to update the setting in the data table with the actual level used in that run. (Follow the advice from @statman)

If you set the levels of a discrete numeric factor to integer values, they should be respected in the design that is created. Are you getting results that disagree with this claim?

[#4]

Thank you for your reply！If I set the numerical variable to discrete type, there will be no problem of decimal in the design table. I'm just not sure whether I can do that. Thank you very much for solving my doubts.