[GeoMechanicsApplication] implement quasi newton strategy

[aronnoordam]

Implement quasi-newton scheme, with options Broyden and/or BFGS.

important is that the user can define where factorization happens with rebuild_level. Either at the beginning of the stage or at the beginning of the solution step. Factorization at every non-linear iteration is not required in this quasi-newton raphson scheme.

Factorization can only be done in direct linear solvers, thus it is important to handle this properly.

Below a scheme of the strategy:

rebuild level 0:

linear_solver->InitializeSolutionStep ----> factorizes rA

for t in range(n_time_steps):
	while not converged:
		update rA_n with broyden

		PerformSolutionStep ---> solves rA u = F

rebuild level 1:

for t in range(n_time_steps):
	linear_solver->InitializeSolutionStep ----> factorizes rA

	while not converged:
		update rA_n with broyden

		PerformSolutionStep ---> solves rA u = F

rebuild level 2 not needed

The method will be benchmarked with the "CROW-case"