Measures the convergence from an old data set to a new one. More...

#include <ResidualRelativeConvergenceMeasure.hpp>

Inheritance diagram for precice::cplscheme::impl::ResidualRelativeConvergenceMeasure:

Collaboration diagram for precice::cplscheme::impl::ResidualRelativeConvergenceMeasure:

Public Member Functions
	ResidualRelativeConvergenceMeasure (double convergenceLimitPercent)
	Constructor.

virtual	~ResidualRelativeConvergenceMeasure ()

virtual void	newMeasurementSeries ()
	To be called when a new meas. series (iteration process) starts.

virtual void	measure (const Eigen::VectorXd &oldValues, const Eigen::VectorXd &newValues)
	Performs convergence measurement.

virtual bool	isConvergence () const
	Returns true, if the last measurement indicates convergence.

virtual std::string	printState (const std::string &dataName)
	Adds current convergence information to output stream.

virtual double	getNormResidual ()
	Returns the l2-norm of the coupling residuum.

virtual std::string	getAbbreviation () const
	Returns an abbreviation of the name of the measure for the log file headers.

Public Member Functions inherited from precice::cplscheme::impl::ConvergenceMeasure
virtual	~ConvergenceMeasure ()
	Destructor, empty.

Private Attributes
logging::Logger	_log {"cplscheme::ResidualRelativeConvergenceMeasure"}

double	_convergenceLimitPercent

bool	_isFirstIteration = true

double	_normFirstResidual = std::numeric_limits<double>::max()

double	_normDiff = 0

bool	_isConvergence = false

Detailed Description

Measures the convergence from an old data set to a new one.

The convergence is evaluated by looking at the two norm of the differences between each data value from the new and old data set. If the two norm is equal or below a given percentage of the norm of the old data set, convergence is achieved.

For a description of how to perform the measurement, see class ConvergenceMeasure.

Definition at line 29 of file ResidualRelativeConvergenceMeasure.hpp.

Constructor & Destructor Documentation

◆ ResidualRelativeConvergenceMeasure()

precice::cplscheme::impl::ResidualRelativeConvergenceMeasure::ResidualRelativeConvergenceMeasure ( double convergenceLimitPercent )

explicit

Constructor.

Parameters

[in] convergenceLimitPercent Limit to define convergence relative to the norm of the current new dataset. Has to be in $] 0 ; 1 ]$.

Definition at line 7 of file ResidualRelativeConvergenceMeasure.cpp.

Here is the call graph for this function:

◆ ~ResidualRelativeConvergenceMeasure()

virtual precice::cplscheme::impl::ResidualRelativeConvergenceMeasure::~ResidualRelativeConvergenceMeasure ( )

inlinevirtual

Definition at line 40 of file ResidualRelativeConvergenceMeasure.hpp.

Member Function Documentation

◆ getAbbreviation()

virtual std::string precice::cplscheme::impl::ResidualRelativeConvergenceMeasure::getAbbreviation ( ) const

inlinevirtual

Returns an abbreviation of the name of the measure for the log file headers.

Reimplemented from precice::cplscheme::impl::ConvergenceMeasure.

Definition at line 91 of file ResidualRelativeConvergenceMeasure.hpp.

◆ getNormResidual()

virtual double precice::cplscheme::impl::ResidualRelativeConvergenceMeasure::getNormResidual ( )

inlinevirtual

Returns the l2-norm of the coupling residuum.

Reimplemented from precice::cplscheme::impl::ConvergenceMeasure.

Definition at line 83 of file ResidualRelativeConvergenceMeasure.hpp.

Here is the call graph for this function:

◆ isConvergence()

virtual bool precice::cplscheme::impl::ResidualRelativeConvergenceMeasure::isConvergence ( ) const

inlinevirtual

Returns true, if the last measurement indicates convergence.

Implements precice::cplscheme::impl::ConvergenceMeasure.

Definition at line 61 of file ResidualRelativeConvergenceMeasure.hpp.

◆ measure()

virtual void precice::cplscheme::impl::ResidualRelativeConvergenceMeasure::measure	(	const Eigen::VectorXd &	oldValues,
		const Eigen::VectorXd &	newValues )