How do you script optimization of JMP formulas and save the results?

[#1]

I have a detailed formula in my JMP table I need to optimize.  I'll need to incrementally vary 3 inputs and let the rest be optimized. 

Is there a way to lock and change the three variables, optimize the others, and save the results for hundreds or thousands of options?

I can do this manually in the prediction profiler, but it becomes unmanageable for the number of cases I'll need.

Thanks!

[#2]

Not sure I understand this correctly, but if you know the equation, can you make it into a formula column and feed your variables into it via 3 separate columns in the same data table? Then you would just have to fill the 3 input columns with whatever incrementally changing numbers you like.

[#3]

Hi. Yes I have a formula column.  I'm wanting to vary the values and then optimize the remaining inputs.  Like optimizing with the desirability function in the prediction profiler, but automated with a script.

[#4]

I suggest that you use the desirability function in the Prediction Profiler. It will find the settings that satisfy the goal represented by this function.

[#5]

Hi. I was doing this manually.  I have three variables that I am changing the importance factor on and then reoptimizing.  I then would append the results to the data table.  I did this for about 30 points manually, but now need hundreds to thousands of points and would like to perform this via a script if possible. 

I saved the desirability function and updated it to include 3 new variables - the importance factors.  I then have the desirability function (saved to my table) and take the values of the three importance values and use each one/sum in the saved desirability function.  What I need to do now is vary the three importance factors and optimize the remaining input variables for steps in each importance factor.  I'm trying to make a three-dimensional pareto frontier.    

[#6]

I think what you're looking for is the Maximize for Each Grid Point option.

Results:

 

[#7]

The Maximize each data point will work thanks!  Right now for my values 0-2 it's doing, 0, 0.5,1,1.5, and 2.  Is there a way for it to take smaller steps?

[#8]

There's no way to get more grid points directly.

However, the grid points for continuous factors are determined using the min/max of the axis on the factor. So, you can adjust the axes and repeat the command. 

This may take you back to your scripting request and the need to create all the combinations of the axes of all your factors but that's probably easier than scripting the grid generation for all the factors for yourself.