Implement degree-1 and dominance reductions.

This seems to still be buggy, since some of the generated instances
are (a lot) bigger than the ones calculated by other solvers.

A list of sizes of solutions to instances generated with the PACE challenge
solver of the team "WeGotYouCovered" and the solutions sizes of this
solver as comments can be found in `solutions`.

The fact that for example `datasets/soc-Slashdot0811.mtx` does not
consider any branches, but generates a solution more than twice as large
as the optimal solution is a strong indicator that this is a problem
with the reductions and not with the branch-and-bound-part.

Also added support for interrupting the calculation with
`SIGINT`/`Ctrl-C` and printing the current best solution.