OPTPP::OptGSS Class Reference

#include <OptGSS.h>

Inheritance diagram for OPTPP::OptGSS:

List of all members.


Public Member Functions
void	setParams ()
	Assign default values to algorithm parameters.
void	setParams (OptGSS_params op)
	Assign values to parameters as specified in argument.
	OptGSS ()
	OptGSS (NLP0 p, GenSetBase g)
	OptGSS (NLP1 p, GenSetBase g)
	OptGSS (NLP0 p, GenSetBase g, NEWMAT::Matrix &M)
	OptGSS (NLP1 p, GenSetBase g, NEWMAT::Matrix &M)
	OptGSS (NLP0 p, GenSetBase g, OptGSS_params op)
	OptGSS (NLP1 p, GenSetBase g, OptGSS_params op)
	OptGSS (NLP0 p, GenSetBase g, NEWMAT::Matrix &M, OptGSS_params op)
	OptGSS (NLP1 p, GenSetBase g, NEWMAT::Matrix &M, OptGSS_params op)
	~OptGSS ()
void	setStepSize (double s)
void	setStepTol (double s)
void	setStepInc (double s)
void	setStepDec (double s)
void	setMaxIter (int s)
void	setFullSearch (bool s)
bool	extras_searched ()
void	setPrintX (bool s)
void	setPrintG (bool s)
void	printCopyright (bool doit)
void	setComputeGrad (bool s)
void	expandStep ()
void	contractStep ()
void	updateX (double &newfX, NEWMAT::ColumnVector &newX)
bool	StopCondition ()
int	StepCondition ()
int	search ()
virtual void	reset ()
virtual void	initOpt ()
virtual void	optimize ()
virtual int	checkConvg ()
int	checkConvg_fcn ()
int	checkConvg_grad ()
virtual void	printStatus (char *msg)
virtual void	printStatus (char *msg)
void	printStatus (char *msg, bool printXc)
	Print the status of optimizer at Xc, but not Xc.
NEWMAT::ColumnVector	computeSearch (NEWMAT::SymmetricMatrix &)
	Print the status of optimizer at Xc, and Xc if 2n arg==true.
virtual void	acceptStep (int k, int step_type)
virtual void	updateModel (int k, int ndim, NEWMAT::ColumnVector x)
virtual void	readOptInput ()
void	printIter (int iter, int bp)
int	getMPIRank ()
int	getMPISize ()
Protected Member Functions
void	printHeader ()
	Prints header of output file before optimization begins.
void	setpid ()
Protected Attributes
NLP0 *	nlp
	Pointer to an NLP0 object.
NLP1 *	nlp1
	Pointer to an NLP1 object.
NEWMAT::ColumnVector	X
	Best (current) point in search.
double	fX
	Value of objective function at X.
NEWMAT::ColumnVector	gX
	Value of gradient at X (when available).
double	fprev
	stores previous best value
double	Delta
	Step-length parameter.
double	Phi
	Steplength expanding parameter (>=1).
double	Theta
	Steplength contracting parameter (0<Theta<ThetaMax<1).
double	Delta_tol
	Convergence tolerance. Algoritms stops when Theta <= Theta_tol.
int	Iter_max
	Upper limit on the number of iterations.
bool	SearchAll
	search type flag (true ==> search on all direction) -- now defunct
bool	computeGrad
	flag to compute gradient after 1st iteration. Used by trustGSS.
GenSetBase *	gset
	Pointer to an Generating Set object.
NEWMAT::Matrix	extras
	Extra search directions.
bool	extras_srched
	True when extra directions have been searched.
bool	printCOPYRIGHT
	if true copyright header is printed before optimization
bool	printXiter
	flag for printing (up to 3) components of X during iterations
bool	printGiter
	flag for printing (up to 3) components of gX during iterations
int	mpi_rank
int	mpi_size

Constructor & Destructor Documentation

OPTPP::OptGSS::OptGSS ( ) [inline]

OPTPP::OptGSS::OptGSS ( NLP0 * p,

GenSetBase * g

) [inline]

Parameters:

p nonlinear problem object

g generating search object

OPTPP::OptGSS::OptGSS ( NLP1 * p,

GenSetBase * g

) [inline]

Parameters:

p nonlinear problem object

g generating search object

OPTPP::OptGSS::OptGSS ( NLP0 * p,

GenSetBase * g,

NEWMAT::Matrix & M

) [inline]

Parameters:

p nonlinear problem object

g generating search object

M matrix with extra search directions

OPTPP::OptGSS::OptGSS ( NLP1 * p,

GenSetBase * g,

NEWMAT::Matrix & M

) [inline]

Parameters:

p nonlinear problem object

g generating search object

M matrix with extra search directions

OPTPP::OptGSS::OptGSS ( NLP0 * p,

GenSetBase * g,

OptGSS_params op

) [inline]

Parameters:

p nonlinear problem object

g generating search object

op GSS algorithmic options

OPTPP::OptGSS::OptGSS ( NLP1 * p,

GenSetBase * g,

OptGSS_params op

) [inline]

Parameters:

p nonlinear problem object

g generating search object

op GSS algorithmic options

OPTPP::OptGSS::OptGSS ( NLP0 * p,

GenSetBase * g,

NEWMAT::Matrix & M,

OptGSS_params op

) [inline]

Parameters:

p nonlinear problem object

g generating search object

M matrix with extra search directions

op GSS algorithmic options

OPTPP::OptGSS::OptGSS ( NLP1 * p,

GenSetBase * g,

NEWMAT::Matrix & M,

OptGSS_params op

) [inline]

Parameters:

p nonlinear problem object

g generating search object

M matrix with extra search directions

op GSS algorithmic options

OPTPP::OptGSS::~OptGSS ( ) [inline]

Member Function Documentation

virtual void OPTPP::OptGSS::acceptStep ( int k,

int step_type

) [inline, virtual]

Note:
Pure virtual functions
Each derived class must define these functions for themselves

Implements OPTPP::OptDirect.

int OPTPP::OptGSS::checkConvg ( ) [virtual]

Reimplemented from OPTPP::OptDirect.

int OPTPP::OptGSS::checkConvg_fcn ( )

int OPTPP::OptGSS::checkConvg_grad ( )

NEWMAT::ColumnVector OPTPP::OptGSS::computeSearch ( NEWMAT::SymmetricMatrix & ) [inline, virtual]

Print the status of optimizer at Xc, and Xc if 2n arg==true.

Implements OPTPP::OptimizeClass.

void OPTPP::OptGSS::contractStep ( ) [inline]

Default Step Decrement Method

void OPTPP::OptGSS::expandStep ( ) [inline]

Default Step Increment Method

bool OPTPP::OptGSS::extras_searched ( ) [inline]

int OPTPP::OptGSS::getMPIRank ( ) [inline]

int OPTPP::OptGSS::getMPISize ( ) [inline]

void OPTPP::OptGSS::initOpt ( ) [virtual]

void OPTPP::OptGSS::optimize ( ) [virtual]

Reimplemented from OPTPP::OptDirect.

void OPTPP::OptGSS::printCopyright ( bool doit ) [inline]

void OPTPP::OptGSS::printHeader ( ) [protected]

Prints header of output file before optimization begins.

void OPTPP::OptGSS::printIter ( int iter,

int bp

)

void OPTPP::OptGSS::printStatus ( char * msg,

bool printXc

)

Print the status of optimizer at Xc, but not Xc.

virtual void OPTPP::OptGSS::printStatus ( char * msg ) [inline, virtual]

Implements OPTPP::OptimizeClass.

virtual void OPTPP::OptGSS::readOptInput ( ) [inline, virtual]

Reimplemented from OPTPP::OptDirect.

void OPTPP::OptGSS::reset ( ) [virtual]

Reset parameter values
Reimplemented from OPTPP::OptDirect.

int OPTPP::OptGSS::search ( )

Search for improved point;

void OPTPP::OptGSS::setComputeGrad ( bool s ) [inline]

void OPTPP::OptGSS::setFullSearch ( bool s ) [inline]

Let the user set the search strategy

void OPTPP::OptGSS::setMaxIter ( int s ) [inline]

Let the user set the max number of iterations
Reimplemented from OPTPP::OptimizeClass.

void OPTPP::OptGSS::setParams ( OptGSS_params op )

Assign values to parameters as specified in argument.

void OPTPP::OptGSS::setParams ( )

Assign default values to algorithm parameters.

void OPTPP::OptGSS::setpid ( ) [inline, protected]

void OPTPP::OptGSS::setPrintG ( bool s ) [inline]

void OPTPP::OptGSS::setPrintX ( bool s ) [inline]

void OPTPP::OptGSS::setStepDec ( double s ) [inline]

Let the user set the step decrement parameter

void OPTPP::OptGSS::setStepInc ( double s ) [inline]

Let the user set the step increment parameter

void OPTPP::OptGSS::setStepSize ( double s ) [inline]

Let the user set the step size

void OPTPP::OptGSS::setStepTol ( double s ) [inline]

Let the user set the step tolerance

int OPTPP::OptGSS::StepCondition ( )

bool OPTPP::OptGSS::StopCondition ( ) [inline]

Default Stopping criterion

virtual void OPTPP::OptGSS::updateModel ( int k,

int ndim,

NEWMAT::ColumnVector x

) [inline, virtual]

Implements OPTPP::OptDirect.

void OPTPP::OptGSS::updateX ( double & newfX,

NEWMAT::ColumnVector & newX

) [inline]

Default update of X and FX with new values

Member Data Documentation

bool OPTPP::OptGSS::computeGrad [protected]

flag to compute gradient after 1st iteration. Used by trustGSS.

double OPTPP::OptGSS::Delta [protected]

Step-length parameter.

double OPTPP::OptGSS::Delta_tol [protected]

Convergence tolerance. Algoritms stops when Theta <= Theta_tol.

NEWMAT::Matrix OPTPP::OptGSS::extras [protected]

Extra search directions.

bool OPTPP::OptGSS::extras_srched [protected]

True when extra directions have been searched.

double OPTPP::OptGSS::fprev [protected]

stores previous best value

Reimplemented from OPTPP::OptimizeClass.

double OPTPP::OptGSS::fX [protected]

Value of objective function at X.

GenSetBase* OPTPP::OptGSS::gset [protected]

Pointer to an Generating Set object.

NEWMAT::ColumnVector OPTPP::OptGSS::gX [protected]

Value of gradient at X (when available).

int OPTPP::OptGSS::Iter_max [protected]

Upper limit on the number of iterations.

int OPTPP::OptGSS::mpi_rank [protected]

int OPTPP::OptGSS::mpi_size [protected]

NLP0* OPTPP::OptGSS::nlp [protected]

Pointer to an NLP0 object.

NLP1* OPTPP::OptGSS::nlp1 [protected]

Pointer to an NLP1 object.

double OPTPP::OptGSS::Phi [protected]

Steplength expanding parameter (>=1).

bool OPTPP::OptGSS::printCOPYRIGHT [protected]

if true copyright header is printed before optimization

bool OPTPP::OptGSS::printGiter [protected]

flag for printing (up to 3) components of gX during iterations

bool OPTPP::OptGSS::printXiter [protected]

flag for printing (up to 3) components of X during iterations

bool OPTPP::OptGSS::SearchAll [protected]

search type flag (true ==> search on all direction) -- now defunct

double OPTPP::OptGSS::Theta [protected]

Steplength contracting parameter (0<Theta<ThetaMax<1).

NEWMAT::ColumnVector OPTPP::OptGSS::X [protected]

Best (current) point in search.

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

Bug Reports OPT++ Developers Copyright Information GNU Lesser General Public License

Documentation, generated by

, last revised August 30, 2006.

OPTPP::OptGSS Class Reference

Public Member Functions

Protected Member Functions

Protected Attributes

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation