thermalfist::Broyden Class Reference

Class implementing the Broyden method to solve a system of non-linear equations. More...

#include <Broyden.h>

Classes
class	BroydenSolutionCriterium
	Sub-class where it is determined whether the required accuracy is achieved in the Broyden's method. More...

Public Member Functions
	Broyden (BroydenEquations *eqs=NULL, BroydenJacobian *jaco=NULL)
	Construct a new Broyden object. More...
virtual	~Broyden (void)
	Destroy the Broyden object. More...
virtual std::vector< double >	Solve (const std::vector< double > &x0, BroydenSolutionCriterium *solcrit=NULL, int max_iterations=MAX_ITERS)
int	Iterations () const
int	MaxIterations () const
double	MaxDifference () const
void	UseNewton (bool flag)
bool	UseNewton () const

Static Public Attributes
static const double	TOL = 1.0E-10
	Default desired solution accuracy. More...
static const int	MAX_ITERS = 200
	Maximum number of Broyden iterations before terminating. More...

Protected Attributes
BroydenEquations *	m_Equations
BroydenJacobian *	m_Jacobian
int	m_Iterations
int	m_MaxIterations
double	m_MaxDifference
bool	m_UseNewton

Detailed Description

Class implementing the Broyden method to solve a system of non-linear equations.

Definition at line 131 of file Broyden.h.

Constructor & Destructor Documentation

thermalfist::Broyden::Broyden ( BroydenEquations *  eqs = NULL, BroydenJacobian *  jaco = NULL  )

inline

Construct a new Broyden object.

Parameters

eqs	A pointer to BroydenEquations object which specifies the equations to be solved.
jaco	A pointer to BroydenJacobian object which calulates the Jacobian matrix. If nullptr is passed then the Jacobian will be computed using finite differences.

Definition at line 197 of file Broyden.h.

virtual thermalfist::Broyden::~Broyden ( void  )

inlinevirtual

Destroy the Broyden object.

Definition at line 203 of file Broyden.h.

Member Function Documentation

int thermalfist::Broyden::Iterations ( ) const

inline

Returns number of Broyden/Newton iterations used to obtain the solution. Only meaningful to call this after the Solve() method was called. If Iterations() equals MaxIterations(), then the convergence was not achieved.

Returns

Number of Broyden/Newton iterations.

Definition at line 229 of file Broyden.h.

double thermalfist::Broyden::MaxDifference ( ) const

inline

Returns

Maximum deviation from zero of any of the equations.

Definition at line 239 of file Broyden.h.

int thermalfist::Broyden::MaxIterations ( ) const

inline

Returns

Maximum number of Broyden/Newton iterations.

Definition at line 234 of file Broyden.h.

std::vector< double > thermalfist::Broyden::Solve ( const std::vector< double > &  x0, BroydenSolutionCriterium *  solcrit = NULL, int  max_iterations = MAX_ITERS  )

virtual

Solves the system of equations. Will use at most max_iterations iterations. Uses the Broyden's method by default, see https://en.wikipedia.org/wiki/Broyden%27s_method Will use the Newton's method https://en.wikipedia.org/wiki/Newton%27s_method if the corresponding flag was set with the UseNewton(bool) method.

Parameters

x0	A vector of starting values of the variables.
solcrit	Criterium used to determine whether the desired accuracy is achieved.
max_iterations	Maximum number of iterations before the Broyden's method terminates.

Returns

std::vector<double> A vector of the variables' values which wolve the equations.

Definition at line 77 of file Broyden.cpp.

void thermalfist::Broyden::UseNewton ( bool  flag )

inline

Specify whether to use Newton's method instead of the Broyden's method. In the Newton's method the Jacobian will be re-evaluated from scrath at each iteration.

Parameters

flag	Use Newton's method if true, use Broyden's method otherwise.

Definition at line 249 of file Broyden.h.

bool thermalfist::Broyden::UseNewton ( ) const

inline

Returns

true Newton's method is to be used

false Broyden's method is to be used

Definition at line 255 of file Broyden.h.

Member Data Documentation

BroydenEquations* thermalfist::Broyden::m_Equations

protected

Definition at line 258 of file Broyden.h.

int thermalfist::Broyden::m_Iterations

protected

Definition at line 260 of file Broyden.h.

BroydenJacobian* thermalfist::Broyden::m_Jacobian

protected

Definition at line 259 of file Broyden.h.

double thermalfist::Broyden::m_MaxDifference

protected

Definition at line 262 of file Broyden.h.

int thermalfist::Broyden::m_MaxIterations

protected

Definition at line 261 of file Broyden.h.

bool thermalfist::Broyden::m_UseNewton

protected

Definition at line 263 of file Broyden.h.

const int thermalfist::Broyden::MAX_ITERS = 200

static

Maximum number of Broyden iterations before terminating.

Definition at line 185 of file Broyden.h.

const double thermalfist::Broyden::TOL = 1.0E-10

static

Default desired solution accuracy.

Definition at line 183 of file Broyden.h.

---

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

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