Improve crossing reduction

[MForster]

Currently the crossing reduction is an extremely simple layer sweep barycenter. This doesn't deal well with many situations, in particular when nodes are only connected in one direction.

I could imagine that iterative improvement steps (trying to re-route individual long) would go a long way. We could also play with the solver that's already embedded anyway.