nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Custom DOE glitch?

cweisbart

I was setting up a DOE with the factors and response from the 20 custom reactor sample files. Instead of going with the default of 20 runs, I was trying to show you could go down to a minimum of 16. When I clicked "Make Design", it instantly went to the page with the design evaluation. Normally, there is an algorithm that runs for about 10 seconds even when I go less than the default, so I thought that was weird. Then when I look at the color map on correlations, everything has an r of 0 (except of course along the diagonal), which does not seem real:

Anyone else getting something like this when going for the minimum amount of runs?

Accepted answers

Phil_Kay

Hi Clovis,
Instantaneous generation of 2-level designs was added in JMP15, I think. And the optimal design result is a standard 2^5-1 for your example, which is resolution V. So there is nothing wrong here that I can see.
Cheers,
Phil

All comments

Phil_Kay

cweisbart

Thanks for the quick response. Ok that absolutely makes sense, I was not thinking. None of the main effects or 2 factor interactions should be aliased with other main effects or 2 factor interactions.

Phil_Kay

Exactly. For the situation that you chose , the optimal solution is a standard design that is orthogonal for all main effects and 2FIs. The design is the same as you would get if you were to use the classical design tools. And the design generation is instantaneous as Custom Design does not need to use the coordinate exchange algorithm for 2 level designs.