#include <opt_interface.hh>

Detailed Description

Definition at line 60 of file opt_interface.hh.

Inheritance diagram for Kaskade::AbstractNormalDirection:

Public Member Functions
void	ordinaryAndAdjoint (AbstractFunctionSpaceElement &correction, AbstractFunctionSpaceElement &adjointCorrection, AbstractLinearization &linearization, AbstractFunctionSpaceElement correctionResidual=nullptr, AbstractFunctionSpaceElement adjointResidual=nullptr)

void	simplified (AbstractFunctionSpaceElement &correction, AbstractLinearization const &linearization, AbstractFunctionSpaceElement *residual=nullptr) const

virtual void	setRelativeAccuracy (double accuracy)

virtual void	setEps (double)

Member Function Documentation

◆ ordinaryAndAdjoint()

void Kaskade::AbstractNormalDirection::ordinaryAndAdjoint	(	AbstractFunctionSpaceElement &	correction,
		AbstractFunctionSpaceElement &	adjointCorrection,
		AbstractLinearization &	linearization,
		AbstractFunctionSpaceElement *	correctionResidual = `nullptr`,
		AbstractFunctionSpaceElement *	adjointResidual = `nullptr`
	)

inline

compute min 1/2 <dn,dn> s.t. c'(x_0)dn+c(x_0)=0 compute Lagrangemultiplier for: min 1/2 <w,w>+f'(x_0) s.t. c'(x_0)w=0 performs factorization uses normal linearization at x_0

Definition at line 67 of file opt_interface.hh.

◆ setEps()

virtual void Kaskade::AbstractNormalDirection::setEps ( double )

inlinevirtual

Reimplemented in Kaskade::PPCGAsNormalSolver< Assembler_, PrecondAssembler, Domain_, Range_, VariableSet, components >.

Definition at line 78 of file opt_interface.hh.

◆ setRelativeAccuracy()

virtual void Kaskade::AbstractNormalDirection::setRelativeAccuracy ( double accuracy )

inlinevirtual

Reimplemented in Kaskade::Bridge::PINVSolver< VectorImpl, InnerSolver >, Kaskade::PPCGAsNormalSolver< Assembler_, PrecondAssembler, Domain_, Range_, VariableSet, components >, and Kaskade::DirectNormalSolver< Assembler_, PrecondAssembler, Domain_, Range_, VariableSet, components >.

Definition at line 76 of file opt_interface.hh.

◆ simplified()

void Kaskade::AbstractNormalDirection::simplified	(	AbstractFunctionSpaceElement &	correction,
		AbstractLinearization const &	linearization,
		AbstractFunctionSpaceElement *	residual = `nullptr`
	)		const

inline

compute min 1/2 <dn,dn> s.t. c'(x_0)dn+c(x)=0 reuses factorization from ordinary(...) or ordinaryAndAdjoint(...) linearization at x

Definition at line 73 of file opt_interface.hh.

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