Maximize output of subvi with two inputs

[fpdx]

I have a subvi which accepts 2 numbers in input and gives one number in output.

What's the best strategy to find the values of the two inputs that maximize the output? Is there an already existing solution?

More info: the subvi is not a mathematical function, but a process that finds and tracks an object in a stream of images. The two inputs "tune" the internal state of the tracker to function properly. The output is an indicator of the goodness of the tracker (the higher the output the better). The goal here is to find the best parameters that make the tracker work the best on each image.

[crossrulz]

Taking me back to my grad school days...

You will need multiple models and inputs.  You will likely want these to run in parallel.  So make your VI reentrant.  This will allow multiple copies of the same VI to run.  From there, use a parallelized FOR loop to run as many models as you have.  Then just choose the maximum one (Array Max & Min).

[Alexander_Sobolev]

Check Mathematics - Optimization pallette (Full LabVIEW development system).

Some of the functions (unconstrained, constrained, global optimization vis) can take the description of your model as a VI, a list of variables to optimize and restricting parameters.

They are universal, so they may not be as efficient as custom solution.