I used RSM for an optimization study using design expert 13.
I expected that the software will generate optimum parameters which may not necessarily be part of the experimental runs as I have seen in most articles. I really dont know how to navigate my way into finding the optimum parameters. Any advise ?
Hi Oyebamiji
How can Design-Expert tell you which factors shall be set to optimum if you're not telling the program where your optimum is in your response? Also you need to tell the program constraints in your factors, otherwise it will be "too easy" for Design-Expert to find desirable solutions.
Probably some factors are more easy to control than others! You have to consider this when proceeding with optimizations.
Basically, you need to understand what comes after RSM, which is the Optimization in DX13. Here you can find a tutorial from statease:
https://www.statease.com/docs/v11/tutorials/multifactor-rsm-pt2/
Another important thing: You never extrapolate! Rather augment your design space to regions where your optimum might be and execute those runs. Your model error is just too big outside the design space to be able to reliably extrapolate.
- Badges
- Science method