Subscribe Bookmark RSS Feed

Asymmetric number of runs for a blocking factor in a custom DOE


Community Member


Sep 18, 2013

I have been trying to generate a DOE for a situation with a strange constraint.  I have a total of 4 categorical variables where one of them is hard to alter with two levels.  One of the constraints that I need to work with is the need to have 80% of the runs at one level of the hard to change factor and then the other 20% of the runs at the other level.  Normally I would say that I need to have 120 total runs and then there would be 4 whole plots which would just put 30 runs per factor level for a total of 60 runs at one level and then 60 runs at the other level.  I realize that I could just generate a larger design and then remove rows one by one but I would like to avoid the need to do that.  Is there a way to get JMP (ver 10+) to do this asymmetric splitting?DOE, cus