Hello Markus and Luca,
Thank you very much for your suggestions and explanation.
It seems to me that both approaches are similar, however, Luca's approach
will allow me to reuse the factorization with different RHSs which, please
correct me if I am mistaken, I might not be able to do using Markus'
approach.
Luca, you mention "Apply boundary values to the matrix, with a fake
boundary_values map, in which all entries which you want to later apply are
set to one, a zero rhs" So, do you mean something like this
BlockVector<double> rhs;
rhs.reinit(solution);
std::map<unsigned int, double> boundary_values;
VectorTools::interpolate_boundary_values (dof,
grid.get_boundary_id("c_CL/Membrane"),
ConstantFunction<dim> PEM(1),
boundary_values,
comp_mask);
MatrixTools::apply_boundary_values(boundary_values,
matrix,
solution,
rhs,
false);
Then you mention "Factorize the matrix, then when you want to apply it to
different bondary conditions, scale the values of the rhs corresponding to
the dirichlet conditions with the previously stored rhs values" What do you
mean by "scale the values of the rhs (...)"?
Thank you,
Marc
_______________________________________________
dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
