How should I set DoE design when there are some operationally dependent parameters ?

[#1]

For my experiment, I have two variables: (1) additive concentration (0, 1, 5 mg/L) and (2) timing for additive supplementation (during cell seeding, during plasmid transfection). Since the timing becomes irrelevant when the additive concentration is 0 mg/L, how could I assign the factors / their levels that account for this operational dependency during designing DoE using JMP ?

[#2]

Hi @Naruchit,

Welcome in the Community !

Looking at your topic, I think you might be interested to look at "Disallowed combinations" option in the Define Factor Constraints panel.

Trying to reproduce your experimental setup with the two factors, if I add a constraint through the Disallowed combinations filter that experiments with an additive concentration close to 0 and a timing higher than 0,01min (or any very low value close to 0) shouldn't be part of the design, I can generate a design corresponding to your needs : 

And here is the resulting design for estimating main effects and the 2-factors interaction :

No experiments in the design have a timing different from 0 when additive concentration is set at 0. 

Here is the script to generate the same design:

DOE(
	Custom Design,
	{Add Response( Maximize, "Y", ., ., . ),
	Add Factor( Discrete Numeric, {0, 1, 5}, "Additive concentration", 0, "mg/L" ),
	Add Factor( Continuous, 0, 10, "Timing", 0, "min" ), Set Random Seed( 98491640 ),
	Number of Starts( 20541 ), Add Term( {1, 0} ), Add Term( {1, 1} ),
	Add Potential Term( {1, 2} ), Add Term( {2, 1} ), Add Term( {1, 1}, {2, 1} ),
	Set Sample Size( 12 ), Disallowed Combinations(
		Additive concentration <= 0.1 & Timing >= 0.01
	), Simulate Responses( 0 ), Save X Matrix( 0 ), Make Design}
);

And you can also find the datatable attached.

If you want to have intermediate level value for timing (like 5min in my example), you can add in the model panel the quadratic term for timing (Timing x Timing). Adding this term will add experiments in the design with an intermediate value for the factor Timing.

Another option in the case of more complex constraints could be to use the candidate set approach, generating a large table of possible experiments, and using it as a candidate set so that the Custom Design platform can choose the most relevant and informative run to create the design.

For more information about factors constraints, I highly recommend reading the blog from @Jed_Campbell Demystifying Factor Constraints. 

Hope this answer will help you,