unitarity constaint

[wkotlarski]

This PR creates a wrapper over SARAH`s unitarity constraint calculation.

How it works:
We create a matrix of functions

matrix[0][0] = [&model,sChan,tChan,uChan,qChan](double sqrtS) {
...
}

each function being one process, e.g. {Ah, hh}->{Ah, hh}. We maximize each entry over $\sqrt{s}$ and then select maximal entry.

TODO:

• 1D function minimization over a bounded region (minimize function within bounds #521)
• loop over generation indices of external particles
• configuration flags