DoE with replicates on the extremities

[#1]

Hello,

I would appreciate support on the following topic:

- I pretend to perform a Full Factorial DoE with 2 center points and replicates on the extremities. When I set up my design I chose 1 replicate (in the tab below) and then the DoE appears one replicate for each experience. Is it possible to set up just one replicate for each one of the extreme conditions? How?

Thank you in advance,

Helena

[#2 · Accepted solution]

Hi @HelenaG,

Depending on the design platform you choose (Custom Design or Classical Screening Design), there is a slight difference in the terms used by JMP for the replication part:
- In a "Classical design", JMP displays "Number of Replicates", meaning here the number of times to perform each run (in addition to the original runs),
- In a "Custom design" JMP displays "Number of Replicate Runs", meaning here the number of run(s) to repeat.

From what I understand from your DoE, the Custom Design platform would better suit you, in order to replicate only some of the runs, and not all the runs.

May I ask you how/why is there a blocking factor in your design ? Is it due to experimental constraints, like 2 experiments per day ?

If you're interested in main effects and interactions for 3 continuous factors with a blocking factor (I don't know how many runs per block you have specified for the blocking factor, but by looking at the design choices it seems you have chosen a block size of 2), you can go into the Custom Design platform, specify your 3 continuous factors, your blocking factor, and in the "Model" part, add all the 2-factors interactions. You can then add 2 centre points and 2 replicate runs (or more, if you want all "extreme" points with 3 factors, that means you may be interested into adding 8 replicate runs, corresponding to the 8 vertices of the 3-dimensional cubic experimental space) :

Since you're in a full factorial design, by default the optimality criterion would be a D-optimal design, so replicate runs will be chosen at the borders of your experimental space (for a more precise estimation of effects coefficients).

Here is the script to generate the design (with 2 centre points and 2 replicate runs) : 

DOE(
	Custom Design,
	{Add Response( Maximize, "Y", ., ., . ),
	Add Factor( Continuous, -1, 1, "X1", 0 ),
	Add Factor( Continuous, -1, 1, "X2", 0 ),
	Add Factor( Continuous, -1, 1, "X3", 0 ), Add Factor( Blocking, 2, "X4" ),
	Set Random Seed( 1327309445 ), Number of Starts( 23862 ), Add Term( {1, 0} ),
	Add Term( {1, 1} ), Add Term( {2, 1} ), Add Term( {3, 1} ), Add Term( {4, 1} ),
	Add Term( {1, 1}, {2, 1} ), Add Term( {1, 1}, {3, 1} ),
	Add Term( {2, 1}, {3, 1} ), Center Points( 2 ), Replicates( 2 ),
	Set Sample Size( 20 ), Simulate Responses( 0 ), Save X Matrix( 0 ), Make Design}
)

And attached you'll find the datatable generated by this script.

Don't hesitate to generate several designs with different number of centre points and replicate runs, in order to compare them and choose the most adequate one depending on your experimental budget and target/needs. 

I hope this first answer will help you,

[#3]

Thank you for the detailed explanation Victor! It was helpful!