Augment DOE with new factor

I would like to be able to add a factor to a previously ran Custom DOE. So an extension of the augment DOE platform that allows me to add a new factor where I could tell you what that factors was set to for the runs that are already done.

Tracking Number:

Defect ID:

5 Comments
Super User

Interesting idea, but I can imagine things would get really interesting if the whole prior experiment is 1 random block, and then you want to fully randomize that factor in the new runs.

Staff (Retired)

I was able to accomplish this by adding a new column and entering the setting for the factor of interest. Then when I augment I include that factor in setup and edit the range in the dialog box. The resulting augment with block has the additional runs at the new setting and codes the column info accordingly. 

Level IV

Yes this can be manually hacked but it would be nice if it was driven by the augment platform and also gave me a recommendation as to how many extra runs I need to accomplish proper characterization.

Staff (Retired)

I would hardly call this "manually hacked" No software can read your mind if you have a factor that you knew the setting of during the initial experiment in question that you later decide to add and include in a subsequent study. Once the new factor is captured the rest is totally driven by the augment platform.

Level VI

If you want to repeat the original design but add a new factor in, you can do this currently by creating a new design, adding the new factor, and they using the original DOE table add the factors as covariates.  

 

If you observed the factor (it was an uncontrolled covariate) in the original design, you can add the covariate column and it's values into the DOE table, and then augment using all original DOE factors and the covariate.