Abstract definition of a System that is used to solve parabolic partial differential equations. More...

#include <OperationParabolicPDESolverSystem.hpp>

Inheritance diagram for sgpp::solver::OperationParabolicPDESolverSystem:

Public Member Functions
void	abortTimestep ()
	aborts the current timestep execution More...

virtual void	coarsenAndRefine (bool isLastTimestep=false)=0

virtual void	finishTimestep ()=0
	performs some action that might be needed after a timestep has be finished in the ODE Solver, e.g. More...

virtual sgpp::base::DataVector *	generateRHS ()=0
	generates the right hand side of the system More...

sgpp::base::Grid *	getGrid ()
	get the pointer to the underlying grid object More...

sgpp::base::DataVector *	getGridCoefficients ()
	gets a pointer to the sparse grids coefficients with evtl. More...

virtual sgpp::base::DataVector *	getGridCoefficientsForCG ()=0
	gets a pointer to the sparse grids coefficients used in the CG method to solve one timestep. More...

void	getGridCoefficientsForSC (sgpp::base::DataVector &Values)
	stores the values of the (dehierarchized) grid in the sgpp::base::DataVector Values used by time step size control methods More...

sgpp::base::GridStorage *	getGridStorage ()

std::string	getODESolver ()
	Returns the specified ODE solver for this instance. More...

sgpp::base::GridStorage *	getOldGridStorage ()

sgpp::base::GridStorage *	getSecondGridStorage ()

size_t	getSumGridPointsComplete ()
	Returns the number of average grid points for the complete grid. More...

size_t	getSumGridPointsInner ()
	Returns the number of average grid points for the inner grid. More...

virtual void	mult (sgpp::base::DataVector &alpha, sgpp::base::DataVector &result)=0
	Multiplicates a vector with the matrix. More...

	OperationParabolicPDESolverSystem ()
	Constructor. More...

void	saveAlpha ()
	stores the current alpha_complete into alpha_complete_old to be available in the next timestep More...

void	setODESolver (std::string ode)
	defines the used ODE Solver for this instance, this is important because the implementation of mult and generateRHS depends on the used ODE solver More...

void	setTimestepSize (double newTimestepSize)
	set the size of the new timestep More...

virtual void	startTimestep ()=0
	Implements some start jobs of every timestep, e.g.discounting boundaries. More...

virtual	~OperationParabolicPDESolverSystem ()
	Destructor. More...

Public Member Functions inherited from sgpp::base::OperationMatrix
	OperationMatrix ()
	Constructor. More...

virtual	~OperationMatrix ()
	Destructor. More...

Protected Attributes
sgpp::base::DataVector *	alpha_complete
	Pointer to the alphas (ansatzfunctions' coefficients) More...

sgpp::base::DataVector *	alpha_complete_old
	Pointer to the alphas from the last timestep, needed when using variable timestep sizes. More...

sgpp::base::DataVector *	alpha_complete_tmp
	Pointer to temporary alphas, needed when using variable timestep sizes. More...

bool	bnewODESolver
	checks whether a new ODE solver has been selected after creation More...

sgpp::base::Grid *	BoundGrid
	Pointer to the grid object. More...

size_t	numSumGridpointsComplete
	Stores number of average gridpoints, complete grid. More...

size_t	numSumGridpointsInner
	Stores number of average gridpoints, inner grid. More...

sgpp::base::GridStorage *	oldGridStorage
	Pointer to the grid from the last iteration. More...

sgpp::base::DataVector *	rhs

sgpp::base::GridStorage *	secondGridStorage
	Pointer to the grid from the last aborted iteration. More...

double	TimestepSize
	the size of one timestep used in the ODE Solver More...

double	TimestepSize_old
	the size of the last timestep More...

std::string	tOperationMode
	specifies in which solver this matrix is used, valid values are: ExEul for explicit Euler ImEul for implicit Euler CrNic for Crank Nicolson solver More...

Detailed Description

Abstract definition of a System that is used to solve parabolic partial differential equations.

So an instance of this class has to pass to any ODE Solver used in SGpp.

\(A \dot{u} = L \vec{u}\)

A: mass matrix L: space discretization (L-Operator)

Constructor & Destructor Documentation

◆ OperationParabolicPDESolverSystem()

sgpp::solver::OperationParabolicPDESolverSystem::OperationParabolicPDESolverSystem ( )

Constructor.

References bnewODESolver, numSumGridpointsComplete, and numSumGridpointsInner.

◆ ~OperationParabolicPDESolverSystem()

sgpp::solver::OperationParabolicPDESolverSystem::~OperationParabolicPDESolverSystem ( )

virtual

Destructor.

Member Function Documentation

◆ abortTimestep()

void sgpp::solver::OperationParabolicPDESolverSystem::abortTimestep ( )

aborts the current timestep execution

References alpha_complete, alpha_complete_tmp, BoundGrid, sgpp::base::Grid::getStorage(), and secondGridStorage.

Referenced by sgpp::solver::VarTimestep::predictor(), sgpp::solver::StepsizeControl::solve(), and sgpp::solver::StepsizeControlH::~StepsizeControlH().

◆ coarsenAndRefine()

virtual void sgpp::solver::OperationParabolicPDESolverSystem::coarsenAndRefine ( bool isLastTimestep = false )

pure virtual

Implemented in sgpp::pde::HeatEquationParabolicPDESolverSystemParallelOMP, and sgpp::pde::HeatEquationParabolicPDESolverSystem.

Referenced by sgpp::solver::AdamsBashforth::solve(), sgpp::solver::CrankNicolson::solve(), sgpp::solver::Euler::solve(), and sgpp::solver::StepsizeControl::solve().

◆ finishTimestep()

virtual void sgpp::solver::OperationParabolicPDESolverSystem::finishTimestep ( )

pure virtual

performs some action that might be needed after a timestep has be finished in the ODE Solver, e.g.

some boundary adjustments.

Implemented in sgpp::pde::HeatEquationParabolicPDESolverSystemParallelOMP, and sgpp::pde::HeatEquationParabolicPDESolverSystem.

Referenced by sgpp::solver::VarTimestep::corrector(), sgpp::solver::VarTimestep::predictor(), sgpp::solver::AdamsBashforth::solve(), sgpp::solver::CrankNicolson::solve(), sgpp::solver::Euler::solve(), sgpp::solver::StepsizeControlEJ::~StepsizeControlEJ(), and sgpp::solver::StepsizeControlH::~StepsizeControlH().

◆ generateRHS()

virtual sgpp::base::DataVector* sgpp::solver::OperationParabolicPDESolverSystem::generateRHS ( )

pure virtual

generates the right hand side of the system

Returns: returns the rhs

Implemented in sgpp::pde::OperationParabolicPDESolverSystemDirichlet, sgpp::pde::HeatEquationParabolicPDESolverSystemParallelOMP, and sgpp::pde::OperationParabolicPDESolverSystemFreeBoundaries.

Referenced by sgpp::solver::VarTimestep::corrector(), sgpp::solver::VarTimestep::predictor(), sgpp::solver::AdamsBashforth::solve(), sgpp::solver::CrankNicolson::solve(), sgpp::solver::Euler::solve(), sgpp::solver::StepsizeControlEJ::~StepsizeControlEJ(), and sgpp::solver::StepsizeControlH::~StepsizeControlH().

◆ getGrid()

sgpp::base::Grid * sgpp::solver::OperationParabolicPDESolverSystem::getGrid ( )

get the pointer to the underlying grid object

Returns: returns a pointer to the underlying grid object

References BoundGrid.

Referenced by sgpp::solver::Euler::solve().

◆ getGridCoefficients()

sgpp::base::DataVector * sgpp::solver::OperationParabolicPDESolverSystem::getGridCoefficients ( )

gets a pointer to the sparse grids coefficients with evtl.

boundaries

Returns: alpha vector of complete grid

References alpha_complete.

Referenced by sgpp::solver::VarTimestep::predictor(), sgpp::solver::Euler::solve(), sgpp::solver::StepsizeControl::solve(), sgpp::solver::StepsizeControlEJ::~StepsizeControlEJ(), and sgpp::solver::StepsizeControlH::~StepsizeControlH().

◆ getGridCoefficientsForCG()

virtual sgpp::base::DataVector* sgpp::solver::OperationParabolicPDESolverSystem::getGridCoefficientsForCG ( )

pure virtual

gets a pointer to the sparse grids coefficients used in the CG method to solve one timestep.

This is useful because (direchlet) boundaries can be skipped when solving the system.

Returns: alpha vector for CG method

Implemented in sgpp::pde::OperationParabolicPDESolverSystemDirichlet, and sgpp::pde::OperationParabolicPDESolverSystemFreeBoundaries.

Referenced by sgpp::solver::VarTimestep::corrector(), sgpp::solver::StepsizeControl::maxNorm(), sgpp::solver::VarTimestep::predictor(), sgpp::solver::AdamsBashforth::solve(), sgpp::solver::CrankNicolson::solve(), sgpp::solver::Euler::solve(), sgpp::solver::StepsizeControlEJ::~StepsizeControlEJ(), and sgpp::solver::StepsizeControlH::~StepsizeControlH().

◆ getGridCoefficientsForSC()

void sgpp::solver::OperationParabolicPDESolverSystem::getGridCoefficientsForSC ( sgpp::base::DataVector & Values )

stores the values of the (dehierarchized) grid in the sgpp::base::DataVector Values used by time step size control methods

Parameters

Values sgpp::base::DataVector in which the values will be stored

References alpha_complete, sgpp::op_factory::createOperationHierarchisation(), and sgpp::base::OperationHierarchisation::doDehierarchisation().

Referenced by sgpp::solver::VarTimestep::corrector(), sgpp::solver::VarTimestep::predictor(), sgpp::solver::StepsizeControl::solve(), sgpp::solver::StepsizeControlEJ::~StepsizeControlEJ(), and sgpp::solver::StepsizeControlH::~StepsizeControlH().

◆ getGridStorage()

sgpp::base::GridStorage * sgpp::solver::OperationParabolicPDESolverSystem::getGridStorage ( )

References BoundGrid, and sgpp::base::Grid::getStorage().

Referenced by sgpp::pde::OperationParabolicPDESolverSystemDirichlet::generateRHS(), and sgpp::solver::StepsizeControl::maxNorm().

◆ getODESolver()

std::string sgpp::solver::OperationParabolicPDESolverSystem::getODESolver ( )

Returns the specified ODE solver for this instance.

Returns: the ODE solver: ExEul, ImEul or CrNic

References tOperationMode.

◆ getOldGridStorage()

sgpp::base::GridStorage * sgpp::solver::OperationParabolicPDESolverSystem::getOldGridStorage ( )

References oldGridStorage.

Referenced by sgpp::pde::OperationParabolicPDESolverSystemDirichlet::generateRHS(), and sgpp::solver::StepsizeControl::maxNorm().

◆ getSecondGridStorage()

sgpp::base::GridStorage * sgpp::solver::OperationParabolicPDESolverSystem::getSecondGridStorage ( )

References secondGridStorage.

◆ getSumGridPointsComplete()

size_t sgpp::solver::OperationParabolicPDESolverSystem::getSumGridPointsComplete ( )

Returns the number of average grid points for the complete grid.

Returns: the number of average grid points for the complete grid

References numSumGridpointsComplete.

◆ getSumGridPointsInner()

size_t sgpp::solver::OperationParabolicPDESolverSystem::getSumGridPointsInner ( )

Returns the number of average grid points for the inner grid.

Returns: the number of average grid points for the inner grid

References numSumGridpointsInner.

◆ mult()

virtual void sgpp::solver::OperationParabolicPDESolverSystem::mult	(	sgpp::base::DataVector &	alpha,
		sgpp::base::DataVector &	result
	)

pure virtual

Multiplicates a vector with the matrix.

Parameters

alpha	sgpp::base::DataVector that contains the ansatzfunctions' coefficients
result	sgpp::base::DataVector into which the result of the space discretization operation is stored

Implements sgpp::base::OperationMatrix.

Implemented in sgpp::pde::OperationParabolicPDESolverSystemDirichlet, sgpp::pde::HeatEquationParabolicPDESolverSystemParallelOMP, and sgpp::pde::OperationParabolicPDESolverSystemFreeBoundaries.

◆ saveAlpha()

void sgpp::solver::OperationParabolicPDESolverSystem::saveAlpha ( )

stores the current alpha_complete into alpha_complete_old to be available in the next timestep

References alpha_complete, alpha_complete_old, alpha_complete_tmp, BoundGrid, sgpp::base::Grid::getStorage(), and oldGridStorage.

Referenced by sgpp::solver::AdamsBashforth::solve(), and sgpp::solver::StepsizeControl::solve().

◆ setODESolver()

void sgpp::solver::OperationParabolicPDESolverSystem::setODESolver ( std::string ode )

defines the used ODE Solver for this instance, this is important because the implementation of mult and generateRHS depends on the used ODE solver

Parameters

ode	the used ODESolver: ExEul, ImEul or CrNic

References bnewODESolver, and tOperationMode.

Referenced by sgpp::solver::VarTimestep::corrector(), sgpp::solver::VarTimestep::predictor(), sgpp::solver::AdamsBashforth::solve(), sgpp::solver::StepsizeControlEJ::~StepsizeControlEJ(), and sgpp::solver::StepsizeControlH::~StepsizeControlH().

◆ setTimestepSize()

void sgpp::solver::OperationParabolicPDESolverSystem::setTimestepSize ( double newTimestepSize )

set the size of the new timestep

Parameters

newTimestepSize the size of the next timestep

References TimestepSize, and TimestepSize_old.

Referenced by sgpp::solver::VarTimestep::predictor(), sgpp::solver::StepsizeControlEJ::~StepsizeControlEJ(), and sgpp::solver::StepsizeControlH::~StepsizeControlH().

◆ startTimestep()

virtual void sgpp::solver::OperationParabolicPDESolverSystem::startTimestep ( )

pure virtual

Implements some start jobs of every timestep, e.g.discounting boundaries.

Implemented in sgpp::pde::HeatEquationParabolicPDESolverSystemParallelOMP, and sgpp::pde::HeatEquationParabolicPDESolverSystem.

Referenced by sgpp::pde::OperationParabolicPDESolverSystemFreeBoundaries::generateRHS(), and sgpp::pde::OperationParabolicPDESolverSystemDirichlet::generateRHS().

Member Data Documentation

◆ alpha_complete

sgpp::base::DataVector* sgpp::solver::OperationParabolicPDESolverSystem::alpha_complete

protected

Pointer to the alphas (ansatzfunctions' coefficients)

◆ alpha_complete_old

sgpp::base::DataVector* sgpp::solver::OperationParabolicPDESolverSystem::alpha_complete_old

protected

Pointer to the alphas from the last timestep, needed when using variable timestep sizes.

Referenced by sgpp::pde::OperationParabolicPDESolverSystemFreeBoundaries::generateRHS(), sgpp::pde::OperationParabolicPDESolverSystemDirichlet::generateRHS(), and saveAlpha().

◆ alpha_complete_tmp

sgpp::base::DataVector* sgpp::solver::OperationParabolicPDESolverSystem::alpha_complete_tmp

protected

Pointer to temporary alphas, needed when using variable timestep sizes.

Referenced by abortTimestep(), and saveAlpha().

◆ bnewODESolver

bool sgpp::solver::OperationParabolicPDESolverSystem::bnewODESolver

protected

checks whether a new ODE solver has been selected after creation

Referenced by OperationParabolicPDESolverSystem(), and setODESolver().

◆ BoundGrid

sgpp::base::Grid* sgpp::solver::OperationParabolicPDESolverSystem::BoundGrid

protected

Pointer to the grid object.

Referenced by abortTimestep(), getGrid(), getGridStorage(), sgpp::pde::HeatEquationParabolicPDESolverSystem::HeatEquationParabolicPDESolverSystem(), sgpp::pde::HeatEquationParabolicPDESolverSystemParallelOMP::HeatEquationParabolicPDESolverSystemParallelOMP(), and saveAlpha().

◆ numSumGridpointsComplete

size_t sgpp::solver::OperationParabolicPDESolverSystem::numSumGridpointsComplete

protected

Stores number of average gridpoints, complete grid.

Referenced by getSumGridPointsComplete(), OperationParabolicPDESolverSystem(), sgpp::pde::OperationParabolicPDESolverSystemDirichlet::OperationParabolicPDESolverSystemDirichlet(), and sgpp::pde::OperationParabolicPDESolverSystemFreeBoundaries::OperationParabolicPDESolverSystemFreeBoundaries().

◆ numSumGridpointsInner

size_t sgpp::solver::OperationParabolicPDESolverSystem::numSumGridpointsInner

protected

Stores number of average gridpoints, inner grid.

Referenced by getSumGridPointsInner(), OperationParabolicPDESolverSystem(), sgpp::pde::OperationParabolicPDESolverSystemDirichlet::OperationParabolicPDESolverSystemDirichlet(), and sgpp::pde::OperationParabolicPDESolverSystemFreeBoundaries::OperationParabolicPDESolverSystemFreeBoundaries().

◆ oldGridStorage

sgpp::base::GridStorage* sgpp::solver::OperationParabolicPDESolverSystem::oldGridStorage

protected

Pointer to the grid from the last iteration.

Referenced by getOldGridStorage(), and saveAlpha().

◆ rhs

sgpp::base::DataVector* sgpp::solver::OperationParabolicPDESolverSystem::rhs

protected

◆ secondGridStorage

sgpp::base::GridStorage* sgpp::solver::OperationParabolicPDESolverSystem::secondGridStorage

protected

Pointer to the grid from the last aborted iteration.

Referenced by abortTimestep(), and getSecondGridStorage().

◆ TimestepSize

double sgpp::solver::OperationParabolicPDESolverSystem::TimestepSize

protected

the size of one timestep used in the ODE Solver

◆ TimestepSize_old

double sgpp::solver::OperationParabolicPDESolverSystem::TimestepSize_old

protected

the size of the last timestep

Referenced by sgpp::pde::OperationParabolicPDESolverSystemFreeBoundaries::generateRHS(), sgpp::pde::OperationParabolicPDESolverSystemDirichlet::generateRHS(), sgpp::pde::OperationParabolicPDESolverSystemFreeBoundaries::mult(), sgpp::pde::OperationParabolicPDESolverSystemDirichlet::mult(), and setTimestepSize().

◆ tOperationMode

std::string sgpp::solver::OperationParabolicPDESolverSystem::tOperationMode

protected

specifies in which solver this matrix is used, valid values are: ExEul for explicit Euler ImEul for implicit Euler CrNic for Crank Nicolson solver

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

solver/src/sgpp/solver/operation/hash/OperationParabolicPDESolverSystem.hpp
solver/src/sgpp/solver/operation/hash/OperationParabolicPDESolverSystem.cpp

Public Member Functions

Protected Attributes

Detailed Description

Constructor & Destructor Documentation

◆ OperationParabolicPDESolverSystem()

◆ ~OperationParabolicPDESolverSystem()

Member Function Documentation

◆ abortTimestep()

◆ coarsenAndRefine()

◆ finishTimestep()

◆ generateRHS()

◆ getGrid()

◆ getGridCoefficients()

◆ getGridCoefficientsForCG()

◆ getGridCoefficientsForSC()

◆ getGridStorage()

◆ getODESolver()

◆ getOldGridStorage()

◆ getSecondGridStorage()

◆ getSumGridPointsComplete()

◆ getSumGridPointsInner()

◆ mult()

◆ saveAlpha()

◆ setODESolver()

◆ setTimestepSize()

◆ startTimestep()

Member Data Documentation

◆ alpha_complete

◆ alpha_complete_old

◆ alpha_complete_tmp

◆ bnewODESolver

◆ BoundGrid

◆ numSumGridpointsComplete

◆ numSumGridpointsInner

◆ oldGridStorage

◆ rhs

◆ secondGridStorage

◆ TimestepSize

◆ TimestepSize_old

◆ tOperationMode