General driver for linear solvers for sparse indefinite symmetric matrices. More...

#include <IpTSymLinearSolver.hpp>

Inheritance diagram for Ipopt::TSymLinearSolver:

Public Member Functions
bool	InitializeImpl (const OptionsList &options, const std::string &prefix)
	Implementation of the initialization method that has to be overloaded by for each derived class.

Constructor/Destructor
	TSymLinearSolver (SmartPtr< SparseSymLinearSolverInterface > solver_interface, SmartPtr< TSymScalingMethod > scaling_method)
	Constructor.

virtual	~TSymLinearSolver ()
	Destructor.

Methods for requesting solution of the linear system.
virtual ESymSolverStatus	MultiSolve (const SymMatrix &A, std::vector< SmartPtr< const Vector > > &rhsV, std::vector< SmartPtr< Vector > > &solV, bool check_NegEVals, Index numberOfNegEVals)
	Solve operation for multiple right hand sides.

virtual Index	NumberOfNegEVals () const
	Number of negative eigenvalues detected during last factorization.


virtual bool	IncreaseQuality ()
	Request to increase quality of solution for next solve.

virtual bool	ProvidesInertia () const
	Query whether inertia is computed by linear solver.

Methods related to the detection of linearly dependent
rows in a matrix
bool	ProvidesDegeneracyDetection () const

ESymSolverStatus	DetermineDependentRows (Index n_rows, Index n_cols, Index n_jac_nz, Number jac_c_vals, Index jac_c_iRow, Index *jac_c_jCol, std::list< Index > &c_deps)

Public Member Functions inherited from Ipopt::SymLinearSolver
	SymLinearSolver ()

virtual	~SymLinearSolver ()

ESymSolverStatus	Solve (const SymMatrix &A, const Vector &rhs, Vector &sol, bool check_NegEVals, Index numberOfNegEVals)
	Solve operation for a single right hand side.

Public Member Functions inherited from Ipopt::AlgorithmStrategyObject
bool	Initialize (const Journalist &jnlst, IpoptNLP &ip_nlp, IpoptData &ip_data, IpoptCalculatedQuantities &ip_cq, const OptionsList &options, const std::string &prefix)
	This method is called every time the algorithm starts again - it is used to reset any internal state.

bool	ReducedInitialize (const Journalist &jnlst, const OptionsList &options, const std::string &prefix)
	Reduced version of the Initialize method, which does not require special Ipopt information.

	AlgorithmStrategyObject ()
	Default Constructor.

virtual	~AlgorithmStrategyObject ()
	Destructor.

Public Member Functions inherited from Ipopt::ReferencedObject
	ReferencedObject ()

virtual	~ReferencedObject ()

Index	ReferenceCount () const

void	AddRef (const Referencer *referencer) const

void	ReleaseRef (const Referencer *referencer) const

Static Public Member Functions

static void	RegisterOptions (SmartPtr< RegisteredOptions > roptions)
	Methods for OptionsList.

Private Member Functions
Default Compiler Generated Methods
(Hidden to avoid implicit creation/calling). These methods are not implemented and we do not want the compiler to implement them for us, so we declare them private and do not define them. This ensures that they will not be implicitly created/called.
	TSymLinearSolver ()
	Default Constructor.

	TSymLinearSolver (const TSymLinearSolver &)
	Copy Constructor.

void	operator= (const TSymLinearSolver &)
	Default Assignment Operator.

Internal functions
ESymSolverStatus	InitializeStructure (const SymMatrix &symT_A)
	Initialize nonzero structure.

void	GiveMatrixToSolver (bool new_matrix, const SymMatrix &sym_A)
	Copy the elements of the matrix in the required format into the array that is provided by the solver interface.

Private Attributes
SmartPtr< SparseSymLinearSolverInterface >	solver_interface_
	Strategy Object for an interface to a linear solver.

Information about the matrix
TaggedObject::Tag	atag_
	Tag for the incoming matrix.

Index	dim_
	Number of rows and columns of the matrix.

Index	nonzeros_triplet_
	Number of nonzeros of the matrix in triplet format.

Index	nonzeros_compressed_
	Number of nonzeros in compressed format.

Initialization flags
bool	have_structure_
	Flag indicating if the internal structures are initialized.

bool	linear_scaling_on_demand_
	Flag indicating whether the scaling objected is to be switched on when increased quality is requested.

bool	initialized_
	Flag indicating if the InitializeStructure method has been called for the linear solver.

Stuff for scaling of the linear system.
SmartPtr< TSymScalingMethod >	scaling_method_
	Strategy Object for a method that computes scaling factors for the matrices.

Number *	scaling_factors_
	Array storing the scaling factors.

bool	use_scaling_
	Flag indicating whether scaling should be performed.

bool	just_switched_on_scaling_
	Flag indicating whether we just switched on the scaling.

information about the matrix.
Index *	airn_
	row indices of matrix in triplet (MA27) format.

Index *	ajcn_
	column indices of matrix in triplet (MA27) format.

SmartPtr< TripletToCSRConverter >	triplet_to_csr_converter_
	Pointer to object for conversion from triplet to compressed format.

SparseSymLinearSolverInterface::EMatrixFormat	matrix_format_
	Flag indicating what matrix data format the solver requires.

Algorithmic parameters
bool	warm_start_same_structure_
	Flag indicating whether the TNLP with identical structure has already been solved before.

Additional Inherited Members
Protected Member Functions inherited from Ipopt::AlgorithmStrategyObject
const Journalist &	Jnlst () const

IpoptNLP &	IpNLP () const

IpoptData &	IpData () const

IpoptCalculatedQuantities &	IpCq () const

bool	HaveIpData () const

Detailed Description

General driver for linear solvers for sparse indefinite symmetric matrices.

This interface includes a call to a method for scaling of the matrix (if given). This class takes in the constructor a pointer to the interface to an actual linear solver, and possibly a pointer to a method for computing scaling factors. It translates the SymMatrix into the format required by the linear solver and calls the solver via the TSymLinearSolverInterface. If a scaling method has been given, the matrix, the right hand side, and the solution are scaled.

Definition at line 33 of file IpTSymLinearSolver.hpp.

Constructor & Destructor Documentation

◆ TSymLinearSolver() [1/3]

Ipopt::TSymLinearSolver::TSymLinearSolver	(	SmartPtr< SparseSymLinearSolverInterface >	solver_interface,
		SmartPtr< TSymScalingMethod >	scaling_method
	)

Constructor.

The solver_interface is a pointer to a linear solver for symmetric matrices in triplet format. If scaling_method not NULL, it must be a pointer to a class for computing scaling factors for the matrix.

◆ ~TSymLinearSolver()

virtual Ipopt::TSymLinearSolver::~TSymLinearSolver ( )

virtual

Destructor.

◆ TSymLinearSolver() [2/3]

Ipopt::TSymLinearSolver::TSymLinearSolver ( )

private

Default Constructor.

◆ TSymLinearSolver() [3/3]

Ipopt::TSymLinearSolver::TSymLinearSolver ( const TSymLinearSolver & )

private

Copy Constructor.

Member Function Documentation

◆ InitializeImpl()

bool Ipopt::TSymLinearSolver::InitializeImpl	(	const OptionsList &	options,
		const std::string &	prefix
	)

virtual

Implementation of the initialization method that has to be overloaded by for each derived class.

Implements Ipopt::SymLinearSolver.

◆ MultiSolve()

virtual ESymSolverStatus Ipopt::TSymLinearSolver::MultiSolve	(	const SymMatrix &	A,
		std::vector< SmartPtr< const Vector > > &	rhsV,
		std::vector< SmartPtr< Vector > > &	solV,
		bool	check_NegEVals,
		Index	numberOfNegEVals
	)

virtual

Solve operation for multiple right hand sides.

Solves the linear system A * Sol = Rhs with multiple right hand sides. If necessary, A is factorized. Correct solutions are only guaranteed if the return values is SYMSOLVER_SUCCESS. The solver will return SYMSOLVER_SINGULAR if the linear system is singular, and it will return SYMSOLVER_WRONG_INERTIA if check_NegEVals is true and the number of negative eigenvalues in the matrix does not match numberOfNegEVals.

check_NegEVals cannot be chosen true, if ProvidesInertia() returns false.

Implements Ipopt::SymLinearSolver.

◆ NumberOfNegEVals()

virtual Index Ipopt::TSymLinearSolver::NumberOfNegEVals ( ) const

virtual

Number of negative eigenvalues detected during last factorization.

Returns: the number of negative eigenvalues of the most recent factorized matrix

This must not be called if the linear solver does not compute this quantities (see ProvidesInertia).

Implements Ipopt::SymLinearSolver.

◆ IncreaseQuality()

virtual bool Ipopt::TSymLinearSolver::IncreaseQuality ( )

virtual

Request to increase quality of solution for next solve.

Ask linear solver to increase quality of solution for the next solve (e.g. increase pivot tolerance).

Returns: false, if this is not possible, e.g. maximal pivot tolerance already used.

Implements Ipopt::SymLinearSolver.

◆ ProvidesInertia()

virtual bool Ipopt::TSymLinearSolver::ProvidesInertia ( ) const

virtual

Query whether inertia is computed by linear solver.

Returns: true, if linear solver provides inertia

Implements Ipopt::SymLinearSolver.

◆ ProvidesDegeneracyDetection()

bool Ipopt::TSymLinearSolver::ProvidesDegeneracyDetection ( ) const

◆ DetermineDependentRows()

ESymSolverStatus Ipopt::TSymLinearSolver::DetermineDependentRows	(	Index	n_rows,
		Index	n_cols,
		Index	n_jac_nz,
		Number *	jac_c_vals,
		Index *	jac_c_iRow,
		Index *	jac_c_jCol,
		std::list< Index > &	c_deps
	)

◆ RegisterOptions()

static void Ipopt::TSymLinearSolver::RegisterOptions ( SmartPtr< RegisteredOptions > roptions )

static

Methods for OptionsList.

◆ operator=()

void Ipopt::TSymLinearSolver::operator= ( const TSymLinearSolver & )

private

Default Assignment Operator.

◆ InitializeStructure()

ESymSolverStatus Ipopt::TSymLinearSolver::InitializeStructure ( const SymMatrix & symT_A )

private

Initialize nonzero structure.

Set dim_ and nonzeros_, and copy the nonzero structure of symT_A into airn_ and ajcn_.

◆ GiveMatrixToSolver()

void Ipopt::TSymLinearSolver::GiveMatrixToSolver	(	bool	new_matrix,
		const SymMatrix &	sym_A
	)

private

Copy the elements of the matrix in the required format into the array that is provided by the solver interface.

Member Data Documentation

◆ atag_

TaggedObject::Tag Ipopt::TSymLinearSolver::atag_

private

Tag for the incoming matrix.

Definition at line 128 of file IpTSymLinearSolver.hpp.

◆ dim_

Index Ipopt::TSymLinearSolver::dim_

private

Number of rows and columns of the matrix.

Definition at line 131 of file IpTSymLinearSolver.hpp.

◆ nonzeros_triplet_

Index Ipopt::TSymLinearSolver::nonzeros_triplet_

private

Number of nonzeros of the matrix in triplet format.

Note that some elements might appear multiple times in which case the values are added.

Definition at line 138 of file IpTSymLinearSolver.hpp.

◆ nonzeros_compressed_

Index Ipopt::TSymLinearSolver::nonzeros_compressed_

private

Number of nonzeros in compressed format.

This is only computed if the sparse linear solver works with the CSR format.

Definition at line 144 of file IpTSymLinearSolver.hpp.

◆ have_structure_

bool Ipopt::TSymLinearSolver::have_structure_

private

Flag indicating if the internal structures are initialized.

For initialization, this object needs to have seen a matrix.

Definition at line 153 of file IpTSymLinearSolver.hpp.

◆ linear_scaling_on_demand_

bool Ipopt::TSymLinearSolver::linear_scaling_on_demand_

private

Flag indicating whether the scaling objected is to be switched on when increased quality is requested.

Definition at line 157 of file IpTSymLinearSolver.hpp.

◆ initialized_

bool Ipopt::TSymLinearSolver::initialized_

private

Flag indicating if the InitializeStructure method has been called for the linear solver.

Definition at line 161 of file IpTSymLinearSolver.hpp.

◆ solver_interface_

SmartPtr<SparseSymLinearSolverInterface> Ipopt::TSymLinearSolver::solver_interface_

private

Strategy Object for an interface to a linear solver.

Definition at line 165 of file IpTSymLinearSolver.hpp.

◆ scaling_method_

SmartPtr<TSymScalingMethod> Ipopt::TSymLinearSolver::scaling_method_

private

Strategy Object for a method that computes scaling factors for the matrices.

If NULL, no scaling is performed.

Definition at line 173 of file IpTSymLinearSolver.hpp.

◆ scaling_factors_

Number* Ipopt::TSymLinearSolver::scaling_factors_

private

Array storing the scaling factors.

Definition at line 175 of file IpTSymLinearSolver.hpp.

◆ use_scaling_

bool Ipopt::TSymLinearSolver::use_scaling_

private

Flag indicating whether scaling should be performed.

Definition at line 177 of file IpTSymLinearSolver.hpp.

◆ just_switched_on_scaling_

bool Ipopt::TSymLinearSolver::just_switched_on_scaling_

private

Flag indicating whether we just switched on the scaling.

Definition at line 179 of file IpTSymLinearSolver.hpp.

◆ airn_

Index* Ipopt::TSymLinearSolver::airn_

private

row indices of matrix in triplet (MA27) format.

Definition at line 185 of file IpTSymLinearSolver.hpp.

◆ ajcn_

Index* Ipopt::TSymLinearSolver::ajcn_

private

column indices of matrix in triplet (MA27) format.

Definition at line 187 of file IpTSymLinearSolver.hpp.

◆ triplet_to_csr_converter_

SmartPtr<TripletToCSRConverter> Ipopt::TSymLinearSolver::triplet_to_csr_converter_

private

Pointer to object for conversion from triplet to compressed format.

This is only required if the linear solver works with the compressed representation.

Definition at line 193 of file IpTSymLinearSolver.hpp.

◆ matrix_format_

SparseSymLinearSolverInterface::EMatrixFormat Ipopt::TSymLinearSolver::matrix_format_

private

Flag indicating what matrix data format the solver requires.

Definition at line 195 of file IpTSymLinearSolver.hpp.

◆ warm_start_same_structure_

bool Ipopt::TSymLinearSolver::warm_start_same_structure_

private

Flag indicating whether the TNLP with identical structure has already been solved before.

Definition at line 203 of file IpTSymLinearSolver.hpp.

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

src/Algorithm/LinearSolvers/IpTSymLinearSolver.hpp

Public Member Functions

Static Public Member Functions

Private Member Functions

Private Attributes

Additional Inherited Members

Detailed Description

Constructor & Destructor Documentation

◆ TSymLinearSolver() [1/3]

◆ ~TSymLinearSolver()

◆ TSymLinearSolver() [2/3]

◆ TSymLinearSolver() [3/3]

Member Function Documentation

◆ InitializeImpl()

◆ MultiSolve()

◆ NumberOfNegEVals()

◆ IncreaseQuality()

◆ ProvidesInertia()

◆ ProvidesDegeneracyDetection()

◆ DetermineDependentRows()

◆ RegisterOptions()

◆ operator=()

◆ InitializeStructure()

◆ GiveMatrixToSolver()

Member Data Documentation

◆ atag_

◆ dim_

◆ nonzeros_triplet_

◆ nonzeros_compressed_

◆ have_structure_

◆ linear_scaling_on_demand_

◆ initialized_

◆ solver_interface_

◆ scaling_method_

◆ scaling_factors_

◆ use_scaling_

◆ just_switched_on_scaling_

◆ airn_

◆ ajcn_

◆ triplet_to_csr_converter_

◆ matrix_format_

◆ warm_start_same_structure_