Implementation of weighted Max-cut

[artix41]

The current implementation of Max-cut only works for non-weighted graphs. However, the more general version can be useful for several applications, like clustering (where the adjacency matrix is the Gram matrix of the data). Qiskit Aqua already has this version for instance.

To implement it, we would just need an adjacency matrix instead of a list of tuples, and assign the given weights to the PauliTerms.