Kaskade::AbstractLinearization Class Reference

Abstract linearization. More...

#include <abstract_interface.hh>

Detailed Description

Abstract linearization.

Given a nonlinear functional, this class represents second order information at the point of linearization

Definition at line 166 of file abstract_interface.hh.

Inheritance diagram for Kaskade::AbstractLinearization:

Public Member Functions
virtual double	eval () const =0
	Evaluate functional. More...
virtual double	evalL1norm () const =0
	Evaluate L1 norm of integrand. More...
virtual void	evald (AbstractFunctionSpaceElement &g, int rbegin=0, int rend=-1) const =0
	Evaluate derivative. More...
void	ddxpy (AbstractFunctionSpaceElement &y, AbstractFunctionSpaceElement const &x, int rbegin=0, int rend=-1, int cbegin=0, int cend=-1) const
	Evaluate hessian times second argument: y = y+ddf*x. More...
void	ddtxpy (AbstractFunctionSpaceElement &y, AbstractFunctionSpaceElement const &x, int rbegin=0, int rend=-1, int cbegin=0, int cend=-1) const
	Evaluate hessian times second argument: y = y+ddf*x. More...
virtual void	d2axpy (double a, AbstractFunctionSpaceElement &y, AbstractFunctionSpaceElement const &x, int rbegin=0, int rend=-1, int cbegin=0, int cend=-1) const =0
	Evaluate hessian times second argument: y = y+a*ddf*x. More...
virtual void	d2taxpy (double a, AbstractFunctionSpaceElement &y, AbstractFunctionSpaceElement const &x, int rbegin=0, int rend=-1, int cbegin=0, int cend=-1) const =0
	Evaluate hessian times second argument: y = y+a*ddf*x. More...
virtual void	getMatrixBlocks (MatrixAsTriplet< double > &mat, int rbegin=0, int rend=-1, int cbegin=0, int cend=-1) const =0
	Access matrix representation if available. More...
virtual AbstractFunctionSpaceElement const &	getOrigin () const =0
	Get point of linearization. More...
virtual void	precompute ()=0
virtual void	flush ()=0
virtual	~AbstractLinearization ()

Constructor & Destructor Documentation

~AbstractLinearization()

virtual Kaskade::AbstractLinearization::~AbstractLinearization ( )

inlinevirtual

Definition at line 198 of file abstract_interface.hh.

Member Function Documentation

d2axpy()

virtual void Kaskade::AbstractLinearization::d2axpy ( double  a, AbstractFunctionSpaceElement &  y, AbstractFunctionSpaceElement const &  x, int  rbegin = 0, int  rend = -1, int  cbegin = 0, int  cend = -1  ) const

pure virtual

Evaluate hessian times second argument: y = y+a*ddf*x.

Implemented in Kaskade::Bridge::KaskadeLinearization< Functional >, and Kaskade::LagrangeLinearization.

Referenced by ddxpy().

d2taxpy()

virtual void Kaskade::AbstractLinearization::d2taxpy ( double  a, AbstractFunctionSpaceElement &  y, AbstractFunctionSpaceElement const &  x, int  rbegin = 0, int  rend = -1, int  cbegin = 0, int  cend = -1  ) const

pure virtual

Evaluate hessian times second argument: y = y+a*ddf*x.

Implemented in Kaskade::Bridge::KaskadeLinearization< Functional >, and Kaskade::LagrangeLinearization.

Referenced by ddtxpy().

ddtxpy()

void Kaskade::AbstractLinearization::ddtxpy ( AbstractFunctionSpaceElement &  y, AbstractFunctionSpaceElement const &  x, int  rbegin = 0, int  rend = -1, int  cbegin = 0, int  cend = -1  ) const

inline

Evaluate hessian times second argument: y = y+ddf*x.

Definition at line 181 of file abstract_interface.hh.

ddxpy()

void Kaskade::AbstractLinearization::ddxpy ( AbstractFunctionSpaceElement &  y, AbstractFunctionSpaceElement const &  x, int  rbegin = 0, int  rend = -1, int  cbegin = 0, int  cend = -1  ) const

inline

Evaluate hessian times second argument: y = y+ddf*x.

Definition at line 176 of file abstract_interface.hh.

eval()

virtual double Kaskade::AbstractLinearization::eval ( ) const

pure virtual

Evaluate functional.

Implemented in Kaskade::Bridge::KaskadeLinearization< Functional >, and Kaskade::LagrangeLinearization.

evald()

virtual void Kaskade::AbstractLinearization::evald ( AbstractFunctionSpaceElement &  g, int  rbegin = 0, int  rend = -1  ) const

pure virtual

Evaluate derivative.

Implemented in Kaskade::LagrangeLinearization, and Kaskade::Bridge::KaskadeLinearization< Functional >.

evalL1norm()

virtual double Kaskade::AbstractLinearization::evalL1norm ( ) const

pure virtual

Evaluate L1 norm of integrand.

Implemented in Kaskade::Bridge::KaskadeLinearization< Functional >, and Kaskade::LagrangeLinearization.

flush()

virtual void Kaskade::AbstractLinearization::flush ( )

pure virtual

Implemented in Kaskade::Bridge::KaskadeLinearization< Functional >, and Kaskade::LagrangeLinearization.

getMatrixBlocks()

virtual void Kaskade::AbstractLinearization::getMatrixBlocks ( MatrixAsTriplet< double > &  mat, int  rbegin = 0, int  rend = -1, int  cbegin = 0, int  cend = -1  ) const

pure virtual

Access matrix representation if available.

Implemented in Kaskade::LagrangeLinearization.

getOrigin()

virtual AbstractFunctionSpaceElement const & Kaskade::AbstractLinearization::getOrigin ( ) const

pure virtual

Get point of linearization.

Implemented in Kaskade::Bridge::KaskadeLinearization< Functional >, and Kaskade::LagrangeLinearization.

Referenced by Kaskade::IdentityChart::addPerturbation(), Kaskade::PrimalChart::addPerturbation(), Kaskade::NormalStepNewton< Newton >::getSearchDirection(), Kaskade::TangentialStepNewton< Newton >::getSearchDirection(), and Kaskade::TangentialStepArmijo::setLinearization().

precompute()

virtual void Kaskade::AbstractLinearization::precompute ( )

pure virtual

Implemented in Kaskade::Bridge::KaskadeLinearization< Functional >, and Kaskade::LagrangeLinearization.

---

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

• abstract_interface.hh