SolverConjugateGradient Class Reference

Conjugate Gradient minimization. More...

#include <numeric_optimization.h>

Inheritance diagram for SolverConjugateGradient:

Public Member Functions
	SolverConjugateGradient (int, ObjectiveScalarInterface &)
	~SolverConjugateGradient (void)
void	solve (double *p)
	returns the location of the minimum More...
void	test (double *x)
	tests the objective function in the neighborhood of the supplied point More...
Public Member Functions inherited from Diagnostic
	Diagnostic (void)
void	setVerbosity (int verbosity)
	sets the verbosity level for this instance More...
int	verbosity (void) const

Public Attributes
int	verbosityLocal
double *	g
double *	h
double *	xi
double *	pcom
double *	xicom
double *	xt
double *	df
int	ITMAX
	the maximum allowed number of iterations More...
double	EPS
	a small number to rectify the special case of converging to exactly zero function value; is a small number that protects against trying to achieve fractional accuracy for a minimum that happens to be exactly zero. More...
double	TOL
	the tolerance passed to dbrent More...
double	GOLD
	the default ratio by which successive intervals are magnified in mnbrak More...
double	GLIMIT
	the maximum magnification allowed for a parabolic-fit step. More...
double	TINY
	used to prevent any possible division by zero in mnbrak_ More...
double	FTOL
	convergence tolerance on the function value More...
int	iter
	the number of iterations that were performed More...
double	fret

Protected Member Functions
double	f1dim_ (double x)
double	df1dim_ (double x)
void	mov3_ (double &a, double &b, double &c, const double d, const double e, const double f)
void	dlinmin_ (double *p)
double	dbrent_ (double, double, double, double, double *)
void	shft3_ (double &a, double &b, double &c, const double d)
void	mnbrak_ (double *, double *, double *, double *, double *, double *)
Protected Member Functions inherited from Diagnostic
virtual	~Diagnostic ()=default

Protected Attributes
int	n_
	number of functions and of unknowns More...
ObjectiveScalarInterface *	ofp_
	class to contain the function to be minimized More...
Protected Attributes inherited from Diagnostic
int	verbosityInstance

Detailed Description

Conjugate Gradient minimization.

#include <libpf/core/numeric_optimization.h>

Remarks

Warning

Thread safe:

Extendable:

Platform dependencies:

Constructor & Destructor Documentation

SolverConjugateGradient()

SolverConjugateGradient::SolverConjugateGradient	(	int	,
		ObjectiveScalarInterface &
	)

~SolverConjugateGradient()

SolverConjugateGradient::~SolverConjugateGradient

(

void

)

Member Function Documentation

dbrent_()

double SolverConjugateGradient::dbrent_ ( double  , double  , double  , double  , double *    )

protected

df1dim_()

double SolverConjugateGradient::df1dim_ ( double  x )

protected

dlinmin_()

void SolverConjugateGradient::dlinmin_ ( double *  p )

protected

f1dim_()

double SolverConjugateGradient::f1dim_ ( double  x )

protected

mnbrak_()

void SolverConjugateGradient::mnbrak_ ( double *  , double *  , double *  , double *  , double *  , double *    )

protected

mov3_()

void SolverConjugateGradient::mov3_ ( double &  a, double &  b, double &  c, const double  d, const double  e, const double  f  )

protected

shft3_()

void SolverConjugateGradient::shft3_ ( double &  a, double &  b, double &  c, const double  d  )

protected

solve()

void SolverConjugateGradient::solve

(

double *

p

)

returns the location of the minimum

test()

void SolverConjugateGradient::test

(

double *

x

)

tests the objective function in the neighborhood of the supplied point

Member Data Documentation

df

double * SolverConjugateGradient::df

EPS

double SolverConjugateGradient::EPS

a small number to rectify the special case of converging to exactly zero function value; is a small number that protects against trying to achieve fractional accuracy for a minimum that happens to be exactly zero.

fret

double SolverConjugateGradient::fret

the minimum value of the function

FTOL

double SolverConjugateGradient::FTOL

convergence tolerance on the function value

g

double* SolverConjugateGradient::g

GLIMIT

double SolverConjugateGradient::GLIMIT

the maximum magnification allowed for a parabolic-fit step.

GOLD

double SolverConjugateGradient::GOLD

the default ratio by which successive intervals are magnified in mnbrak

h

double * SolverConjugateGradient::h

iter

int SolverConjugateGradient::iter

the number of iterations that were performed

ITMAX

int SolverConjugateGradient::ITMAX

the maximum allowed number of iterations

n_

int SolverConjugateGradient::n_

protected

number of functions and of unknowns

ofp_

ObjectiveScalarInterface* SolverConjugateGradient::ofp_

protected

class to contain the function to be minimized

pcom

double* SolverConjugateGradient::pcom

TINY

double SolverConjugateGradient::TINY

used to prevent any possible division by zero in mnbrak_

TOL

double SolverConjugateGradient::TOL

the tolerance passed to dbrent

verbosityLocal

int SolverConjugateGradient::verbosityLocal

xi

double * SolverConjugateGradient::xi

xicom

double * SolverConjugateGradient::xicom

xt

double* SolverConjugateGradient::xt

---

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

• numeric_optimization.h