QUESO::OperatorBase Class Reference

Abstract base class for operator objects. Operators are assumed to be symmetric and positive-definite. More...

#include <OperatorBase.h>

Public Member Functions
virtual double	get_eigenvalue (unsigned int i) const =0
	Return eigenvalue i. More...
virtual double	get_inverted_eigenvalue (unsigned int i) const =0
	Return the reciprocal of eigenvalue i. More...
virtual unsigned int	get_num_converged () const =0
	Return the number of converged eigenpairs. More...
virtual boost::shared_ptr < FunctionBase >	inverse_kl_transform (std::vector< double > &xi, double alpha) const =0
	Given coefficients xi, computes the Karhunen-Loeve transform. More...
Constructor/Destructor methods
	OperatorBase ()
	Constructor. More...
virtual	~OperatorBase ()
	Destructor. More...

Detailed Description

Abstract base class for operator objects. Operators are assumed to be symmetric and positive-definite.

Definition at line 45 of file OperatorBase.h.

Constructor & Destructor Documentation

QUESO::OperatorBase::OperatorBase

(

)

Constructor.

Definition at line 30 of file OperatorBase.C.

{
}

QUESO::OperatorBase::~OperatorBase ( )

virtual

Destructor.

Definition at line 34 of file OperatorBase.C.

{
}

Member Function Documentation

virtual double QUESO::OperatorBase::get_eigenvalue ( unsigned int  i ) const

pure virtual

Return eigenvalue i.

You can store them however you want, but having some kind of order to them is useful for InfiniteDimensionalMeasure

Referenced by QUESO::InfiniteDimensionalGaussian::get_kl_coefficient().

virtual double QUESO::OperatorBase::get_inverted_eigenvalue ( unsigned int  i ) const

pure virtual

Return the reciprocal of eigenvalue i.

virtual unsigned int QUESO::OperatorBase::get_num_converged ( ) const

pure virtual

Return the number of converged eigenpairs.

Referenced by QUESO::InfiniteDimensionalGaussian::draw(), and QUESO::InfiniteDimensionalGaussian::InfiniteDimensionalGaussian().

virtual boost::shared_ptr<FunctionBase> QUESO::OperatorBase::inverse_kl_transform ( std::vector< double > &  xi, double  alpha  ) const

pure virtual

Given coefficients xi, computes the Karhunen-Loeve transform.

This transform goes from coefficient space to physical space using this as the precision operator: / pow(, alpha / 2.0) (x) where the lambda are eigenvalues of the precision operator, this, and the (x) are eigenfunctions of the precision operator, this

Referenced by QUESO::InfiniteDimensionalGaussian::draw().

---

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

• src/core/inc/OperatorBase.h
• src/core/src/OperatorBase.C