GIMLI::PolynomialModelling Class Reference

Multidimensional polynomial modelling class. More...

Inheritance diagram for GIMLI::PolynomialModelling:

Collaboration diagram for GIMLI::PolynomialModelling:

Public Member Functions
	PolynomialModelling (uint dim, uint nCoeffizient, const std::vector< RVector3 > &referencePoints, const RVector &startModel)
virtual RVector	response (const RVector &par)
virtual RVector	startModel ()
const PolynomialFunction< double > &	polynomialFunction () const
void	setPascalsStyle (bool is)
void	setSerendipityStyle (bool is)
void	setPowCombinationTmp (int i)
Public Member Functions inherited from GIMLI::ModellingBase
	ModellingBase (bool verbose=false)
	ModellingBase (DataContainer &dataContainer, bool verbose=false)
	ModellingBase (const Mesh &mesh, bool verbose=false)
	ModellingBase (const Mesh &mesh, DataContainer &dataContainer, bool verbose=false)
void	setVerbose (bool verbose)
bool	verbose () const
virtual RVector	response_mt (const RVector &model, Index i=0) const
RVector	operator() (const RVector &model)
void	setData (DataContainer &data)
DataContainer &	data () const
virtual RVector	createDefaultStartModel ()
virtual void	setStartModel (const RVector &startModel)
void	setMesh (const Mesh &mesh, bool ignoreRegionManager=false)
Mesh *	mesh ()
void	createRefinedForwardMesh (bool refine=true, bool pRefine=false)
void	deleteMesh ()
virtual void	setJacobian (MatrixBase *J)
virtual void	createJacobian (const RVector &model)
virtual void	createJacobian (const RVector &model, const RVector &resp)
virtual void	createJacobian_mt (const RVector &model, const RVector &resp)
virtual void	initJacobian ()
MatrixBase *	jacobian ()
MatrixBase *	jacobian () const
virtual RMatrix &	jacobianRef () const
virtual RMatrix &	jacobianRef ()
virtual void	clearJacobian ()
RVector	createMappedModel (const RVector &model, double background=-9e99) const
void	setRegionManager (RegionManager *reg)
const RegionManager &	regionManager () const
RegionManager &	regionManager ()
RegionManager &	regionManagerRef ()
bool	verbose ()
Region *	region (int marker)
RVector	createStartModel ()
RVector	createStartVector ()
void	initRegionManager ()
void	setThreadCount (Index nThreads)
Index	threadCount ()
void	setMultiThreadJacobian (Index nThreads)
Index	multiThreadJacobian () const

Protected Attributes
uint	dim_
std::vector< RVector3 >	referencePoints_
PolynomialFunction< double >	f_
bool	pascalTriangle_
bool	serendipityStyle_
uint	powCombination_
Protected Attributes inherited from GIMLI::ModellingBase
Mesh *	mesh_
DataContainer *	dataContainer_
MatrixBase *	jacobian_
bool	ownJacobian_
MatrixBase *	constraints_
bool	ownConstraints_
RMatrix	solutions_
RVector	startModel_
bool	verbose_
bool	regionManagerInUse_
bool	ownRegionManager_
Index	nThreads_
Index	nThreadsJacobian_

Additional Inherited Members
Protected Member Functions inherited from GIMLI::ModellingBase
virtual void	init_ ()
virtual void	deleteMeshDependency_ ()
virtual void	updateMeshDependency_ ()
virtual void	updateDataDependency_ ()
void	setMesh_ (const Mesh &mesh, bool update=true)

Detailed Description

Multidimensional polynomial modelling class.

Multidimensional polynomial modelling class. 1, 2 or 3 dimensional polynomial modelling operator based on PolynomialFunction

Member Function Documentation

polynomialFunction()

const PolynomialFunction< double >& GIMLI::PolynomialModelling::polynomialFunction ( ) const

inline

Return read only reference to the Polynomial Functor.

setPascalsStyle()

void GIMLI::PolynomialModelling::setPascalsStyle ( bool  is )

inline

Constrain to polynomial function based on Pascal's triangle, i.e., forbid x^iy^jz^k with (i +j +k) > size

setSerendipityStyle()

void GIMLI::PolynomialModelling::setSerendipityStyle ( bool  is )

inline

Constrain the polynomial function serendipity style Pascal's triangle, i.e., forbid x^iy^jz^k with (i +j +k) > size+1. Just work if setPascalsStyle is set.

startModel()

RVector GIMLI::PolynomialModelling::startModel ( )

virtual

Create starting model. The dimension is recognized here
one-dimensional: p[0][0][*] = 1
two-dimensional: p[0][*][*] = 1
three-dimensional: p[*][*][*] = 1

Reimplemented from GIMLI::ModellingBase.

References GIMLI::Vector< ValueType >::setVal(), and GIMLI::PolynomialFunction< ValueType >::size().