Factor upper limit change in DOE

[#1]

Hi,

we  have made DOE with 5 factors. At the initial run we noticed that the in one factor the higher value was too high. we made a decision to lower the value before continuing with the following  runs. is it ok to change this value in the other 4 runs by hand or should I redesign the hole DOE from the factors definition step etc.? 

[#2 · Accepted solution]

Hi @YanivD,

From your situation, I understand that for each upper limit in one factor in the DoE, you reduce its value ?
Depending on your DoE, it may have consequences (or not).

• If the factor is continuous and you only evaluate main effects (+ possible interactions with other effects) but no other non-linear effect and no centre points, then upper and lower value are already coded in the DoE as +1 and -1, so I don't think it will change the design.
• If the factor is not continuous and having more than two levels (or continuous with quadratic or higher non linear effects studied) and/or centre points, then you might lose some optimality, as your centre or middle points in your design won't be at the middle of the range because of this upper value change.

One option to check would be to look at the "Evaluate Design" platform of your original design and compare it with your modified design, on the Power Analysis, Prediction Variance Profile and Design diagnostics panels.
If you have the same values/performances, no consequences of this change value on your design :)

I hope this answer will help you,