I've run a DOE on a system with two process factors (time & temperature) and 1 material factor (reagent A level). My goal is to determine how the reagent A level affects the range of time and temperature that can be used to create an acceptable result. So far, I have used the desirability functions for each response to create a desirability formula that can be profiled, as shown below for a particular reagent A level:
Essentially I want to be able to compute a "desirability volume" under that green shape but above some threshold desirability (e.g. 0.5) that will represent a viable application space. An alternative calculation would be the average desirability above the target value.
What would be the best way to go about doing this?