I also like to point out in a few similar contexts: even if you could measure all the things, and you could create a perfect/optimal cost function, there would

I also like to point out in a few similar contexts: even if you could measure all the things, and you could create a perfect/optimal cost function, there would still be no guarantee that you'd be able to find the optimal global solution. So relax.