I think you have picked particularly bad examples for this. Both the model and the target function have real, practical considerations that make them probably impossible. If, however, you did make such a model and such a function then the running that model in every possible configuration space would be a problem of such massive scale that you couldn't ever hope to find the best configuration.
I think if you have some sort of point you are trying to make, you should confine yourself to something more practical and easier to have a definite answer. If you so I think you will find that, eventually, the invalidated belief will be dropped except by a small minority and the consensus will shift to encompass that viewpoint.
I think if you have some sort of point you are trying to make, you should confine yourself to something more practical and easier to have a definite answer. If you so I think you will find that, eventually, the invalidated belief will be dropped except by a small minority and the consensus will shift to encompass that viewpoint.