SparseQR.h file
Sparse QR factorization with column pivoting.
This header provides the SparseQR solver class, which implements QR decomposition with numerical column pivoting for sparse matrices.
Algorithm
The solver computes the following quantity.
where is a permutation matrix, is orthogonal, and is upper triangular. The left-looking algorithm proceeds column by column.
Applicability
- General sparse matrices (no symmetry required)
- Rank-deficient systems
- Least-squares problems
- Overdetermined systems
- Systems where LU factorization is numerically unstable
Usage Example
Problem problem(u, v); problem = Integral(Grad(u), Grad(v)) - Integral(f, v); Solver::SparseQR solver(problem); solver.solve();
Namespaces
- namespace Rodin
- The Rodin library for finite element methods and shape optimization.
- namespace Rodin::FormLanguage
- Typed expressions for constructing variational forms.
- namespace Rodin::Solver
- Solution of linear and nonlinear discrete systems.
Classes
-
template<class LinearSystem>struct Rodin::FormLanguage::Traits<Solver::SparseQR<LinearSystem>>
- Form-language traits for SparseQR solvers.
-
template<class Scalar>class Rodin::Solver::SparseQR<Math::LinearSystem<Math::SparseMatrix<Scalar>, Math::Vector<Scalar>>>
- Sparse QR factorization with numerical column pivoting.