The hybrid smoother by R. Hiptmair. More...

#include <hiptmair_hybrid_smoother.h>

Inheritance diagram for ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >:

Classes
struct	t_red_op_and
	"and" reduction operation class for the conductivity condition More...

struct	t_red_op_or
	"or" reduction operation class for the conductivity condition More...

struct	tEdgeInfo
	Structure of the storage for the information about the edge-vertex interconnections. More...

Public Types
typedef MatrixOperator< matrix_type, vector_type >	matrix_operator_type
	Matrix Operator type. More...

typedef TAlgebra::matrix_type	matrix_type
	Matrix type. More...

typedef MatrixOperator< pot_matrix_type, pot_vector_type >	pot_matrix_operator_type
	Matrix Operator type for the potential space. More...

typedef TPotAlgebra::matrix_type	pot_matrix_type
	Matrix type for the potential space. More...

typedef TPotAlgebra::vector_type	pot_vector_type
	Vector type for the potential space. More...

typedef GridFunction< TDomain, TAlgebra >	TGridFunc
	Grid function type for the solution. More...

typedef CPUAlgebra	TPotAlgebra
	The auxiliary algebra type for the space of the potential. (Note: It should be scalar.) More...

typedef GridFunction< TDomain, TPotAlgebra >	TPotGridFunc
	Grid function type for the potential. More...

typedef TAlgebra::vector_type	vector_type
	Vector type. More...

Public Types inherited from ug::ILinearIterator< class, class >
typedef Y	codomain_function_type

typedef X	domain_function_type

Public Types inherited from ug::DebugWritingObject< TAlgebra >
typedef TAlgebra	algebra_type

typedef TAlgebra::matrix_type	matrix_type

typedef TAlgebra::vector_type	vector_type

Public Types inherited from VectorDebugWritingObject< TAlgebra::vector_type >
typedef TAlgebra::vector_type	vector_type

Public Member Functions
bool	apply (vector_type &c, const vector_type &d)
	Computes the correction. More...

bool	apply_update_defect (vector_type &c, vector_type &d)
	Computes the correction and updates the defect d := d - L*c. More...

SmartPtr< ILinearIterator< vector_type, vector_type > >	clone ()
	Clone the smoother by copying the data. More...

bool	init (SmartPtr< ILinearOperator< vector_type > > J, const vector_type &u)
	Initialization using a matrix and a GridFunction (not merely a vector!) More...

bool	init (SmartPtr< ILinearOperator< vector_type > > L)
	We cannot initialize without the geometry (this version of init cannot not work) More...

const char *	name () const
	Returns the name. More...

void	set_Dirichlet (SmartPtr< EMDirichlet< TDomain, TAlgebra > > spDirichlet)
	Sets the Dirichlet boundary. More...

void	set_skip_edge_smoother (bool skip_edge)
	Skip flag the edge smoother. More...

void	set_skip_vertex_smoother (bool skip_vertex)
	Skip flag the vertex smoother. More...

bool	supports_parallel () const
	Currently returns false because the computation of the potentials is not purely parallel up to now. More...

	TimeHarmonicNedelecHybridSmoother (SmartPtr< ApproximationSpace< TDomain > > vertApproxSpace, SmartPtr< ILinearIterator< vector_type > > edgeSmoother, SmartPtr< ILinearIterator< pot_vector_type > > vertSmoother)
	Constructor setting the approx. spaces and the default subsmoothers. More...

	~TimeHarmonicNedelecHybridSmoother ()
	Destructor. More...

Public Member Functions inherited from ug::ILinearIterator< class, class >
virtual bool	apply (Y &c, const X &d)=0

virtual bool	apply_update_defect (Y &c, X &d)=0

virtual std::string	config_string () const

SmartPtr< IDamping< X, Y > >	damping ()

	ILinearIterator ()

	ILinearIterator (const ILinearIterator< X, Y > &parent)

virtual bool	init (SmartPtr< ILinearOperator< Y, X > > J, const Y &u)=0

virtual bool	init (SmartPtr< ILinearOperator< Y, X > > L)=0

void	set_damp (number factor)

void	set_damp (SmartPtr< IDamping< X, Y > > spScaling)

virtual	~ILinearIterator ()

Public Member Functions inherited from ug::DebugWritingObject< TAlgebra >
SmartPtr< IDebugWriter< algebra_type > >	debug_writer ()

ConstSmartPtr< IDebugWriter< algebra_type > >	debug_writer () const

bool	debug_writer_valid () const

	DebugWritingObject ()

	DebugWritingObject (const DebugWritingObject< algebra_type > &parent)

	DebugWritingObject (SmartPtr< IDebugWriter< algebra_type > > spDebugWriter)

virtual void	set_debug (SmartPtr< IDebugWriter< algebra_type > > spDebugWriter)

virtual	~DebugWritingObject ()

Public Member Functions inherited from VectorDebugWritingObject< TAlgebra::vector_type >
virtual void	set_debug (SmartPtr< IVectorDebugWriter< vector_type > > spDebugWriter)

SmartPtr< IVectorDebugWriter< vector_type > >	vector_debug_writer ()

ConstSmartPtr< IVectorDebugWriter< vector_type > >	vector_debug_writer () const

bool	vector_debug_writer_valid () const

	VectorDebugWritingObject ()

	VectorDebugWritingObject (SmartPtr< IVectorDebugWriter< vector_type > > spDebugWriter)

void	write_debug (const vector_type &vec, const char *filename)

virtual	~VectorDebugWritingObject ()

Private Types
typedef TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >	this_type

Private Member Functions
void	collect_edge_defect (const vector_type &d, pot_vector_type &potDefRe, pot_vector_type &potDefIm)
	Computes the vertex-centered defect \( d_{pot} = G^T d \). More...

void	compute_GtMG ()
	Computes the product \( G^T M^{(1)}_h G \). More...

void	compute_potential_matrix (const DoFDistribution pEdgeDD, const DoFDistribution pVertDD)
	Computes the matrix for the smoother in the potential space and marks the "conductive nodes". More...

void	distribute_vertex_correction (pot_vector_type &potCorRe, pot_vector_type &potCorIm, vector_type &c)
	Adds the vertex-centered correction to the edge-centered one: More...

void	get_edge_vert_correspondence (const DoFDistribution pEdgeDD, const DoFDistribution pVertDD)
	Gets the correspondence between the edges and the vertices. More...

bool	zero_mass_entry (double val)
	Measure for numerically zero entries in the mass matrix. More...

Private Attributes
bool	m_bInit
	Whether initialized. More...

bool	m_bSkipEdge
	Needed mainly for debugging: Whether to skip one of the stages. More...

bool	m_bSkipVertex

GridLevel	m_GridLevel
	Grid level (or surface) where to smooth. More...

TPotGridFunc *	m_pPotCorIm
	Vertex-centered grid function for the Im-part of potential corrections (this is a GridFunction to allow geometry-dependent smoothers) More...

TPotGridFunc *	m_pPotCorRe
	Vertex-centered grid function for the Re-part of potential corrections (this is a GridFunction to allow geometry-dependent smoothers) More...

SmartPtr< EMDirichlet< TDomain, TAlgebra > >	m_spDirichlet
	Dirichlet boundary. More...

SmartPtr< DoFDistribution >	m_spEdgeDD
	DoF distribution for the Nedelec elements. More...

SmartPtr< ILinearIterator< vector_type > >	m_spEdgeSmoother
	Smoother for the edge dofs. More...

SmartPtr< pot_matrix_operator_type >	m_spPotMat
	Vertex-centered matrix for the potential. More...

SmartPtr< matrix_operator_type >	m_spSysMat
	Edge-centered matrix of the original equation. More...

SmartPtr< ApproximationSpace< TDomain > >	m_spVertApproxSpace
	Approximation space for the potential (vertex-centered) space. More...

SmartPtr< DoFDistribution >	m_spVertDD
	Level DoF distribution for the vertex centered grid func. More...

SmartPtr< ILinearIterator< pot_vector_type > >	m_spVertSmoother
	Smoother for the vertex dofs. More...

VariableArray1< tEdgeInfo >	m_vEdgeInfo
	Storage for the information about the edge-vertex interconnections. More...

Additional Inherited Members
Protected Member Functions inherited from ug::DebugWritingObject< TAlgebra >
void	enter_debug_writer_section (const char *secDir)

void	enter_debug_writer_section (std::string secDir)

void	leave_debug_writer_section ()

void	write_debug (const matrix_type &mat, const char *filename)

void	write_debug (const matrix_type &mat, std::string name)

Protected Member Functions inherited from VectorDebugWritingObject< TAlgebra::vector_type >
void	enter_vector_debug_writer_section (const char *secDir)

void	enter_vector_debug_writer_section (std::string secDir)

void	leave_vector_debug_writer_section ()

void	print_debugger_message (const char *msg)

void	print_debugger_message (std::string msg)

virtual void	write_debug (const vector_type &vec, std::string name)

Protected Attributes inherited from ug::ILinearIterator< class, class >
SmartPtr< IDamping< X, Y > >	m_spDamping

Protected Attributes inherited from ug::DebugWritingObject< TAlgebra >
SmartPtr< IDebugWriter< algebra_type > >	m_spDebugWriter

Protected Attributes inherited from VectorDebugWritingObject< TAlgebra::vector_type >
SmartPtr< IVectorDebugWriter< vector_type > >	m_spVectorDebugWriter

Detailed Description

template<typename TDomain, typename TAlgebra>
class ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >

The hybrid smoother by R. Hiptmair.

This class implements the hybrid smoother proposed by R. Hiptmair for the discretizations based on the Nedelec element. The smoother is based on the approximate Helmholtz decomponstion of the operator. This implementation is constructed for the discretizations of the time-harmonic equations. The class computes the matrix for the discrete potential and calls two further smoothers specified by the user: one in the space of the edge elements (Whitney-1 or Nedelec forms) and one in the vertex-centered potential space (i.e. the Whitney-0 or Lagrange-1 elements). Accordingly, this smoother requires 2 approximation spaces: one for the edge dofs and one for the vertex dofs.

References:

R. Hiptmair. Multigrid method for Maxwell’s equations. SIAM J. Numer. Anal., 36(1): pp. 204–225 (1998), DOI 10.1137/S0036142997326203.
R. Beck, P. Deuflhard, R. Hiptmair, R. Hoppe und B. Wohlmuth. Adaptive multilevel methods for edge element discretizations of Maxwell’s equations. Surveys on Mathematics for Industry, 8(3-4): pp. 271–312 (1999).

Template Parameters

TDomain	Type of Domain
TAlgebra	Type of Algebra

Member Typedef Documentation

◆ matrix_operator_type

template<typename TDomain , typename TAlgebra >

typedef MatrixOperator<matrix_type, vector_type> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::matrix_operator_type

Matrix Operator type.

◆ matrix_type

template<typename TDomain , typename TAlgebra >

typedef TAlgebra::matrix_type ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::matrix_type

Matrix type.

◆ pot_matrix_operator_type

template<typename TDomain , typename TAlgebra >

typedef MatrixOperator<pot_matrix_type, pot_vector_type> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::pot_matrix_operator_type

Matrix Operator type for the potential space.

◆ pot_matrix_type

template<typename TDomain , typename TAlgebra >

typedef TPotAlgebra::matrix_type ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::pot_matrix_type

Matrix type for the potential space.

◆ pot_vector_type

template<typename TDomain , typename TAlgebra >

typedef TPotAlgebra::vector_type ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::pot_vector_type

Vector type for the potential space.

◆ TGridFunc

template<typename TDomain , typename TAlgebra >

typedef GridFunction<TDomain, TAlgebra> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::TGridFunc

Grid function type for the solution.

◆ this_type

template<typename TDomain , typename TAlgebra >

typedef TimeHarmonicNedelecHybridSmoother<TDomain,TAlgebra> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::this_type

private

◆ TPotAlgebra

template<typename TDomain , typename TAlgebra >

typedef CPUAlgebra ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::TPotAlgebra

The auxiliary algebra type for the space of the potential. (Note: It should be scalar.)

◆ TPotGridFunc

template<typename TDomain , typename TAlgebra >

typedef GridFunction<TDomain, TPotAlgebra> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::TPotGridFunc

Grid function type for the potential.

◆ vector_type

template<typename TDomain , typename TAlgebra >

typedef TAlgebra::vector_type ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::vector_type

Vector type.

Constructor & Destructor Documentation

◆ TimeHarmonicNedelecHybridSmoother()

template<typename TDomain , typename TAlgebra >

ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::TimeHarmonicNedelecHybridSmoother	(	SmartPtr< ApproximationSpace< TDomain > >	vertApproxSpace,
		SmartPtr< ILinearIterator< vector_type > >	edgeSmoother,
		SmartPtr< ILinearIterator< pot_vector_type > >	vertSmoother
	)

inline

Constructor setting the approx. spaces and the default subsmoothers.

Parameters

vertApproxSpace	vertex-centered approx. space
edgeSmoother	the edge-centered smoother
vertSmoother	the vertex-centered smoother

References SmartPtr< class, FreePolicy >::invalid(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spEdgeSmoother, ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spVertApproxSpace, ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spVertSmoother, ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::name(), and UG_THROW.

◆ ~TimeHarmonicNedelecHybridSmoother()

template<typename TDomain , typename TAlgebra >

ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::~TimeHarmonicNedelecHybridSmoother ( )

inline

Destructor.

References ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_pPotCorIm, and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_pPotCorRe.

Member Function Documentation

◆ apply()

template<typename TDomain , typename TAlgebra >

bool ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::apply	(	vector_type &	c,
		const vector_type &	d
	)

Computes the correction.

This method applies the iterator operator, i.e. invertes the approximate matrix. The defect d remains unchanged.

The idea is to apply first the edge-centered smoother (specified as the argument of the constructor of the class). To the correction obtained in this way, the vertex-centered correction is added. The latter is obtained by the application of the vertex-centered smoother to the system

\begin{eqnarray*} \left ( \begin{array} {cc} 0 & A_{pot} \\ A_{pot} & 0 \end{array} \right ) \left ( \begin{array} {c} \mathbf{Re} c_{pot} \\ \mathbf{Im} c_{pot} \end{array} \right ) = \left ( \begin{array} {c} - \mathbf{Re} d_{pot} \\ \mathbf{Im} d_{pot} \end{array} \right ), \end{eqnarray*}

where \( A_{pot} = G^T M^{(1)}_h G \) and \( d_{pot} = G^T d \). To add the vertex-centered correction, obtained in this way, to the edge-centered one, one sets: \( c := c_{edge} + G c_{pot} \).

Parameters

[in]	d	defect
[out]	c	correction

Returns: bool success flag

References ug::PST_ADDITIVE, ug::PST_CONSISTENT, ParallelVector< Vector< double > >::set_layouts(), ParallelVector< Vector< double > >::set_storage_type(), UG_LOG, and UG_THROW.

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::apply_update_defect().

◆ apply_update_defect()

template<typename TDomain , typename TAlgebra >

bool ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::apply_update_defect	(	vector_type &	c,
		vector_type &	d
	)

inline

Computes the correction and updates the defect d := d - L*c.

References ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::apply(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spSysMat, ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::name(), and UG_THROW.

◆ clone()

template<typename TDomain , typename TAlgebra >

SmartPtr<ILinearIterator<vector_type, vector_type> > ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::clone ( )

inlinevirtual

Clone the smoother by copying the data.

Implements ug::ILinearIterator< class, class >.

References ug::ILinearIterator< class, class >::damping(), ug::DebugWritingObject< TAlgebra >::debug_writer(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_bSkipEdge, ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_bSkipVertex, ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spDirichlet, ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spEdgeSmoother, ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spVertApproxSpace, and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spVertSmoother.

◆ collect_edge_defect()

template<typename TDomain , typename TAlgebra >

void ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::collect_edge_defect	(	const vector_type &	d,
		pot_vector_type &	potDefRe,
		pot_vector_type &	potDefIm
	)

private

Computes the vertex-centered defect \( d_{pot} = G^T d \).

Computes the defect in the potential space: \( d_{pot} = G^T d \), where \( d \) is the defect of the original system. This is done only for the 'conductive' vertices. At the other vertices, \( d_{pot} \) is set to 0.

Parameters

d	original (edge-centered) defect
potDefRe	real part of \( d_{pot} \)
potDefIm	imaginary part of \( d_{pot} \)

References ug::BlockRef(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::conductive_vrt_0(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::conductive_vrt_1(), ParallelVector< Vector< double > >::set(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::vrt_index.

◆ compute_GtMG()

template<typename TDomain , typename TAlgebra >

void ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::compute_GtMG

private

Computes the product \( G^T M^{(1)}_h G \).

Performs the multiplication \( G^T M^{(1)}_h G \). Furthermore, marks the 'conductive' vertices, i.e. the vertices that lie on the closure of the conductive subset (i.e. fills the m_vConductiveVertex array). 'Conductive' vertices are ends of the edges whose entries of the mass matrix \( M^{(1)}_h\) are non-zero. Any other vertex, which is not an end of such an edge, is considered as 'non-conductive'.

For the non-conductive nodes, the function sets the row of the potential stiffness matrix to the identity.

This function uses the m_vEdgeInfo-array to get the correspondence between the edges.

References ug::VariableArray1< class >::at(), ug::BlockRef(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::is_Dirichlet(), ug::VariableArray1< class >::resize(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::set_conductive_vrt_0(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::set_conductive_vrt_1(), ug::SetDirichletRow(), UG_THROW, and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::vrt_index.

◆ compute_potential_matrix()

template<typename TDomain , typename TAlgebra >

void ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::compute_potential_matrix	(	const DoFDistribution *	pEdgeDD,
		const DoFDistribution *	pVertDD
	)

private

Computes the matrix for the smoother in the potential space and marks the "conductive nodes".

Computes the matrix \(A_{pot}\) for the smoother in the potential space and marks the "conductive nodes" for the computation of the correction.

\begin{eqnarray*} A_{pot} = G^T M^{(1)}_h G, \end{eqnarray*}

where \(G\) is the note-to-edge gradient incidence matrix and \(M^{(1)}_h\) is the mass matrix of the Whitney-1-forms, taken from the (2, 1)-position of the block-form of the system matrix of the time-harmonic system:

\begin{eqnarray*} \left ( \begin{array} {cc} - S_h (\mu^{-1}) & \omega M^{(1)}_h (\sigma) \\ \omega M^{(1)}_h (\sigma) & S_h (\mu^{-1}) \end{array} \right ) \end{eqnarray*}

(this is the edge-centered matrix).

Template Parameters

TDomain	type of domain in the class
TAlgebra	type of algebra in the class

Parameters

pEdgeDD	edge-centered DoF distribution of the grid functions
pVertDD	vertex-centered DoF distribution of the grid functions

References ug::DoFDistribution::layouts(), ug::DoFDistribution::num_indices(), UG_CATCH_THROW, and UG_THROW.

◆ distribute_vertex_correction()

template<typename TDomain , typename TAlgebra >

void ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::distribute_vertex_correction	(	pot_vector_type &	potCorRe,
		pot_vector_type &	potCorIm,
		vector_type &	c
	)

private

Adds the vertex-centered correction to the edge-centered one:

Updates the edge-centered correction: \( c := c_{edge} + G c_{pot} \).

Parameters

potCorRe	real part of the potential correction \( c_{pot} \)
potCorIm	imaginary part of the potential correction \( c_{pot} \)
c	final (edge-centered) correction

References ug::BlockRef(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::vrt_index.

◆ get_edge_vert_correspondence()

template<typename TDomain , typename TAlgebra >

void ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::get_edge_vert_correspondence	(	const DoFDistribution *	pEdgeDD,
		const DoFDistribution *	pVertDD
	)

private

Gets the correspondence between the edges and the vertices.

Gets the correspondence between the edge DoFs and the vertex DoFs (i.e. fills the tEdgeInfo-structures in m_vEdgeInfo). Furthermore, this function marks edges at Dirichlet boundaries.

Template Parameters

TDomain	type of domain in the class
TAlgebra	type of algebra in the class

Parameters

pEdgeDD	edge-centered DoF distribution of the grid functions
pVertDD	vertex-centered DoF distribution of the grid functions

References ug::DoFDistribution::begin(), ug::DoFDistribution::end(), ug::DoFDistribution::inner_algebra_indices(), ug::DoFDistribution::num_indices(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::set_init(), ug::SubsetGroup::size(), ug::DoFDistributionInfoProvider::subset_handler(), UG_THROW, ug::EdgeVertices::vertex(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::tEdgeInfo::vrt_index.

◆ init() [1/2]

template<typename TDomain , typename TAlgebra >

bool ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::init	(	SmartPtr< ILinearOperator< vector_type > >	J,
		const vector_type &	u
	)

Initialization using a matrix and a GridFunction (not merely a vector!)

This function initializes the smoother with the matrix J and the geometry read from the GridFunction u. Note that u should be a GridFunction, not merely a vector.

Parameters

[in]	J	the matrix of the time-harmonic Nedelec discretization
[in]	u	the grid function containing the grid info (the values are not used)

Returns: bool success flag

References ug::PST_CONSISTENT, and UG_THROW.

◆ init() [2/2]

template<typename TDomain , typename TAlgebra >

bool ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::init ( SmartPtr< ILinearOperator< vector_type > > L )

inline

We cannot initialize without the geometry (this version of init cannot not work)

References ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::name(), and UG_THROW.

◆ name()

template<typename TDomain , typename TAlgebra >

const char* ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::name ( ) const

inlinevirtual

Returns the name.

Implements ug::ILinearIterator< class, class >.

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::apply_update_defect(), ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::init(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::TimeHarmonicNedelecHybridSmoother().

◆ set_Dirichlet()

template<typename TDomain , typename TAlgebra >

void ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::set_Dirichlet ( SmartPtr< EMDirichlet< TDomain, TAlgebra > > spDirichlet )

inline

Sets the Dirichlet boundary.

Parameters

spDirichlet the Dirichlet BC object

References ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spDirichlet.

◆ set_skip_edge_smoother()

template<typename TDomain , typename TAlgebra >

void ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::set_skip_edge_smoother ( bool skip_edge )

inline

Skip flag the edge smoother.

References ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_bSkipEdge.

◆ set_skip_vertex_smoother()

template<typename TDomain , typename TAlgebra >

void ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::set_skip_vertex_smoother ( bool skip_vertex )

inline

Skip flag the vertex smoother.

References ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_bSkipVertex.

◆ supports_parallel()

template<typename TDomain , typename TAlgebra >

bool ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::supports_parallel ( ) const

inlinevirtual

Currently returns false because the computation of the potentials is not purely parallel up to now.

Implements ug::ILinearIterator< class, class >.

◆ zero_mass_entry()

template<typename TDomain , typename TAlgebra >

bool ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::zero_mass_entry ( double val )

inlineprivate

Measure for numerically zero entries in the mass matrix.

Member Data Documentation

◆ m_bInit

template<typename TDomain , typename TAlgebra >

bool ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_bInit

private

Whether initialized.

◆ m_bSkipEdge

template<typename TDomain , typename TAlgebra >

bool ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_bSkipEdge

private

Needed mainly for debugging: Whether to skip one of the stages.

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::clone(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::set_skip_edge_smoother().

◆ m_bSkipVertex

template<typename TDomain , typename TAlgebra >

bool ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_bSkipVertex

private

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::clone(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::set_skip_vertex_smoother().

◆ m_GridLevel

template<typename TDomain , typename TAlgebra >

GridLevel ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_GridLevel

private

Grid level (or surface) where to smooth.

◆ m_pPotCorIm

template<typename TDomain , typename TAlgebra >

TPotGridFunc* ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_pPotCorIm

private

Vertex-centered grid function for the Im-part of potential corrections (this is a GridFunction to allow geometry-dependent smoothers)

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::~TimeHarmonicNedelecHybridSmoother().

◆ m_pPotCorRe

template<typename TDomain , typename TAlgebra >

TPotGridFunc* ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_pPotCorRe

private

Vertex-centered grid function for the Re-part of potential corrections (this is a GridFunction to allow geometry-dependent smoothers)

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::~TimeHarmonicNedelecHybridSmoother().

◆ m_spDirichlet

template<typename TDomain , typename TAlgebra >

SmartPtr<EMDirichlet<TDomain, TAlgebra> > ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spDirichlet

private

Dirichlet boundary.

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::clone(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::set_Dirichlet().

◆ m_spEdgeDD

template<typename TDomain , typename TAlgebra >

SmartPtr<DoFDistribution> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spEdgeDD

private

DoF distribution for the Nedelec elements.

◆ m_spEdgeSmoother

template<typename TDomain , typename TAlgebra >

SmartPtr<ILinearIterator<vector_type> > ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spEdgeSmoother

private

Smoother for the edge dofs.

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::clone(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::TimeHarmonicNedelecHybridSmoother().

◆ m_spPotMat

template<typename TDomain , typename TAlgebra >

SmartPtr<pot_matrix_operator_type> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spPotMat

private

Vertex-centered matrix for the potential.

◆ m_spSysMat

template<typename TDomain , typename TAlgebra >

SmartPtr<matrix_operator_type> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spSysMat

private

Edge-centered matrix of the original equation.

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::apply_update_defect().

◆ m_spVertApproxSpace

template<typename TDomain , typename TAlgebra >

SmartPtr<ApproximationSpace<TDomain> > ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spVertApproxSpace

private

Approximation space for the potential (vertex-centered) space.

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::clone(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::TimeHarmonicNedelecHybridSmoother().

◆ m_spVertDD

template<typename TDomain , typename TAlgebra >

SmartPtr<DoFDistribution> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spVertDD

private

Level DoF distribution for the vertex centered grid func.

◆ m_spVertSmoother

template<typename TDomain , typename TAlgebra >

SmartPtr<ILinearIterator<pot_vector_type> > ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_spVertSmoother

private

Smoother for the vertex dofs.

Referenced by ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::clone(), and ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::TimeHarmonicNedelecHybridSmoother().

◆ m_vEdgeInfo

template<typename TDomain , typename TAlgebra >

VariableArray1<tEdgeInfo> ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >::m_vEdgeInfo

private

Storage for the information about the edge-vertex interconnections.

The documentation for this class was generated from the following files:

/home/runner/work/docs/docs/ug4/plugins/Electromagnetism/EddyCurrent_E_Nedelec/hiptmair_hybrid_smoother.h
/home/runner/work/docs/docs/ug4/plugins/Electromagnetism/EddyCurrent_E_Nedelec/hiptmair_hybrid_smoother_impl.h

Classes

Public Types

Public Member Functions

Private Types

Private Member Functions

Private Attributes

Additional Inherited Members

Detailed Description

template<typename TDomain, typename TAlgebra> class ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >

Member Typedef Documentation

◆ matrix_operator_type

◆ matrix_type

◆ pot_matrix_operator_type

◆ pot_matrix_type

◆ pot_vector_type

◆ TGridFunc

◆ this_type

◆ TPotAlgebra

◆ TPotGridFunc

◆ vector_type

Constructor & Destructor Documentation

◆ TimeHarmonicNedelecHybridSmoother()

◆ ~TimeHarmonicNedelecHybridSmoother()

Member Function Documentation

◆ apply()

◆ apply_update_defect()

◆ clone()

◆ collect_edge_defect()

◆ compute_GtMG()

◆ compute_potential_matrix()

◆ distribute_vertex_correction()

◆ get_edge_vert_correspondence()

◆ init() [1/2]

◆ init() [2/2]

◆ name()

◆ set_Dirichlet()

◆ set_skip_edge_smoother()

◆ set_skip_vertex_smoother()

◆ supports_parallel()

◆ zero_mass_entry()

Member Data Documentation

◆ m_bInit

◆ m_bSkipEdge

◆ m_bSkipVertex

◆ m_GridLevel

◆ m_pPotCorIm

◆ m_pPotCorRe

◆ m_spDirichlet

◆ m_spEdgeDD

◆ m_spEdgeSmoother

◆ m_spPotMat

◆ m_spSysMat

◆ m_spVertApproxSpace

◆ m_spVertDD

◆ m_spVertSmoother

◆ m_vEdgeInfo

template<typename TDomain, typename TAlgebra>
class ug::Electromagnetism::TimeHarmonicNedelecHybridSmoother< TDomain, TAlgebra >