## Optimize desirability with fixed factors

Solved
Occasional Contributor
Posts: 13

# Optimize desirability with fixed factors

Hi,

I am buiding a DoE model using 6 responses and 1 continuous and 5 categorical factors.

One of the factors is a tablet, I have 6 different.

The other factors are methods, each applicable to a tablet.

I need to find the optimum value of the factors for all tablets, as well as for each of the tablets.

How do I optimise desirability, for each of the tablets? In the prediction profiler it seems I can't fix a factor. The optimal setting can very well be different for each tablet.

Accepted Solutions
Solution
‎10-15-2013 11:03 AM
Occasional Contributor
Posts: 13

## Re: Optimize desirability with fixed factors

It is actually very easy: In the prediction profile grid you Alt-click on the factor. A little window appears, where it is possible to select "Lock Factor Setting".

Then the factor is fixed, and the dashed red line becomes solid. When "Maximise desirability" is then chosen from the red triangle menu this factor is not varied.

More than one factor can be locked.

An alternative solution is to select "Maximise for each grid point" in the red-triangle menu. Then all combinations of desirability is shown. You can then sort after the "desirability" column, to see which setting is best.

All Replies
Solution
‎10-15-2013 11:03 AM
Occasional Contributor
Posts: 13

## Re: Optimize desirability with fixed factors

It is actually very easy: In the prediction profile grid you Alt-click on the factor. A little window appears, where it is possible to select "Lock Factor Setting".

Then the factor is fixed, and the dashed red line becomes solid. When "Maximise desirability" is then chosen from the red triangle menu this factor is not varied.

More than one factor can be locked.

An alternative solution is to select "Maximise for each grid point" in the red-triangle menu. Then all combinations of desirability is shown. You can then sort after the "desirability" column, to see which setting is best.

🔒 This topic is solved and locked.