 | ArnoldiEigenvalueSolver Class |
Solve the generalized eigenvalue problem
Ax = Mx(lambda)
Where A is sparse symmetric and M is sparse symmetric semi position
definite. Solve is accomplished using a shift and invert spectral
transformation and implicitly restarted Arnoldi iteration.
Inheritance HierarchyNamespace: CenterSpace.NMath.Core
Assembly: NMath (in NMath.dll) Version: 7.4
SyntaxThe ArnoldiEigenvalueSolver type exposes the following members.
Constructors| Name | Description | |
|---|---|---|
 | ArnoldiEigenvalueSolver |
Constructs a SparseGeneralizedEigServer instance.
|
Methods| Name | Description | |
|---|---|---|
 | Solve(DoubleSymCsrSparseMatrix, DoubleSymCsrSparseMatrix, ArnoldiEigenvalueOptions) |
Solve the generalized symmetric eigenvalue problem
Ax = Mx(lambda) using Arnoldi iteration,
Where A is symmetric and M is symmetric positive semi-definite.
|
| Solve(DoubleSymmetricMatrix, DoubleSymmetricMatrix, ArnoldiEigenvalueOptions) |
Solve the generalized egivenvalue problem
Ax = Mx(lambda)
Where A is sparse symmetric and M is sparse symmetric semi position
definite. Solve is accomplished using a shift and invert spectral
transformation and implicitly restarted Arnoldi iteration.
|
Remarks
The shift and invert problem is
inv(A - (sigma)M)Mx = x(nu),
where nu = 1/(lambda - sigma).
The transformation is effective for finding eigenvalues near sigma.
See Also