thermalfist::BroydenJacobian Class Reference

Class which implements calculation of the Jacobian needed for the Broyden's method. More...

#include <Broyden.h>

Public Member Functions
	BroydenJacobian (BroydenEquations *eqs=NULL)
	Construct a new BroydenJacobian object. More...
virtual	~BroydenJacobian (void)
	Destructor. More...
virtual std::vector< double >	Jacobian (const std::vector< double > &x)
	Evaluates the Jacobian for given values of the variables. More...
void	SetDx (double dx)
	Set the finite variable difference value used for calculating the Jacobian numerically. More...
double	CurrentDx () const

Static Public Attributes
static const double	EPS = 1.0E-6

Detailed Description

Class which implements calculation of the Jacobian needed for the Broyden's method.

Here it is done numerically using the BroydenEquations instance and the finite difference method. A derived class may implement an analytic calculation.

Definition at line 76 of file Broyden.h.

Constructor & Destructor Documentation

thermalfist::BroydenJacobian::BroydenJacobian ( BroydenEquations *  eqs = NULL )

inline

Construct a new BroydenJacobian object.

Parameters

eqs	A pointer to BroydenEquations object which implements the equations to be solved by the Broyden's method.

Definition at line 89 of file Broyden.h.

virtual thermalfist::BroydenJacobian::~BroydenJacobian ( void  )

inlinevirtual

Destructor.

Definition at line 92 of file Broyden.h.

Member Function Documentation

double thermalfist::BroydenJacobian::CurrentDx ( ) const

inline

Returns

double Current finite variable difference value.

Definition at line 119 of file Broyden.h.

std::vector< double > thermalfist::BroydenJacobian::Jacobian ( const std::vector< double > &  x )

virtual

Evaluates the Jacobian for given values of the variables.

Parameters

x	Vector of the variables' values.

Returns

std::vector<double> The computed Jacobian matrix. Returns a vector of matrix elements stored in a RowMajor ordering.

Here finite differences are used to approximates the derivatives. This method can be overriden in a derived class e.g. to implement analytic calculations for a specific system of equations.

Definition at line 27 of file Broyden.cpp.

void thermalfist::BroydenJacobian::SetDx ( double  dx )

inline

Set the finite variable difference value used for calculating the Jacobian numerically.

Parameters

dx	The finite difference value.

Definition at line 114 of file Broyden.h.

Member Data Documentation

const double thermalfist::BroydenJacobian::EPS = 1.0E-6

static

The default finite variable difference value used for calculations the Jacobian numerically.

Definition at line 81 of file Broyden.h.

---

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

• include/HRGBase/Broyden.h
• src/library/HRGBase/Broyden.cpp