OPTPP::OptBCEllipsoid Class Reference

#include <OptBCEllipsoid.h>

Inheritance diagram for OPTPP::OptBCEllipsoid:

List of all members.


Public Member Functions
	OptBCEllipsoid ()
	OptBCEllipsoid (NLP1 *p)
	OptBCEllipsoid (NLP1 *p, UPDATEFCN u)
	OptBCEllipsoid (NLP1 *p, TOLS t)
	~OptBCEllipsoid ()
virtual void	acceptStep (int k, int step_type)
virtual int	checkConvg ()
	Check to see if algorithm satisfies the convergence criteria.
NEWMAT::ColumnVector	computeSearch (NEWMAT::SymmetricMatrix &)
virtual void	optimize ()
	Call the optimization method.
virtual void	readOptInput ()
	Read user-specified input options.
virtual void	printStatus (char *)
	Print status of the bound constrained ellipsoidal method.
virtual void	updateModel (int k, int ndim, NEWMAT::ColumnVector x)
virtual void	reset ()
virtual void	initOpt ()
	Sets up the optimization method.
void	setInitialEllipsoid (double rad)
	Sets the initial ellipsoid radius.
double	computeFeasibility (NEWMAT::ColumnVector &)
	Computes feasibility of the constraints.
NEWMAT::ColumnVector	computeConstraintSubgradient (NEWMAT::ColumnVector &)
	Picks the row corresponding to the most infeasible constraints.
int	infeasibilityStep (NEWMAT::ColumnVector &, NEWMAT::SymmetricMatrix &, double &)
	Step taken if the current x is infeasible.
int	halfSpaceStep (NEWMAT::ColumnVector &, NEWMAT::SymmetricMatrix &, double &)
	Deep cut step for upper bound.
void	setXScale (NEWMAT::ColumnVector &x)
	Set the scaling vector for x.
void	setDeepCut ()
	Set deepcutflag = 1.
void	resetDeepCut ()
	Set deepcutflag = 0.
double	computeGamma (double)
	Given x, compute the gamma function.
Protected Member Functions
NLP1 *	nlprob () const
	Pointer to an NLP1*.
Protected Attributes
double	fval_lowbound
	Function lower/upper bounds.
double	fval_upbound
	Function lower/upper bounds.
double	initial_radius
	Radius of the initial sphere.
int	xscal_flag
	Scaling indicator.
int	deepcutflag
	Deep cut indicator.
Private Attributes
NLP1 *	mem_nlp
	Pointer to an NLP1*.
int	nlp_type
	NLP identifier.

Detailed Description

Bound Constrained Newton abstract data classes. OptBCEllipsoid implements the bound constrained ellipsoid method.

Author:: J.C. Meza, Sandia National Laboratories, meza@ca.sandia.gov
Charles Tong

Note:: Modified by P.J. Williams to accept new constrained nlp objects

Constructor & Destructor Documentation

OPTPP::OptBCEllipsoid::OptBCEllipsoid ( ) [inline]

Default constructor

OPTPP::OptBCEllipsoid::OptBCEllipsoid ( NLP1 * p ) [inline]

Parameters:

p a pointer to an NLP1 object

See also:
OptBCEllipsoid(NLP1* p, UPDATEFCN u)
OptBCEllipsoid(NLP1* p, TOLS t)

OPTPP::OptBCEllipsoid::OptBCEllipsoid ( NLP1 * p,

UPDATEFCN u

) [inline]

Parameters:

p a pointer to an NLP1 object

u an update function

See also:
OptBCEllipsoid(NLP1* p)
OptBCEllipsoid(NLP1* p, UPDATEFCN u)

OPTPP::OptBCEllipsoid::OptBCEllipsoid ( NLP1 * p,

TOLS t

) [inline]

Parameters:

p a pointer to an NLP1 object

t a TOLS object

See also:
OptBCEllipsoid(NLP1* p)
OptBCEllipsoid(NLP1* p, UPDATEFCN u)

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

Destructor

Member Function Documentation

void OPTPP::OptBCEllipsoid::acceptStep ( int k,

int step_type

) [virtual]

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

Implements OPTPP::OptimizeClass.

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

Check to see if algorithm satisfies the convergence criteria.

Implements OPTPP::OptimizeClass.

ColumnVector OPTPP::OptBCEllipsoid::computeConstraintSubgradient ( NEWMAT::ColumnVector & )

Picks the row corresponding to the most infeasible constraints.

double OPTPP::OptBCEllipsoid::computeFeasibility ( NEWMAT::ColumnVector & )

Computes feasibility of the constraints.

double OPTPP::OptBCEllipsoid::computeGamma ( double )

Given x, compute the gamma function.

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

Implements OPTPP::OptimizeClass.

int OPTPP::OptBCEllipsoid::halfSpaceStep ( NEWMAT::ColumnVector & ,

NEWMAT::SymmetricMatrix & ,

double &

)

Deep cut step for upper bound.

int OPTPP::OptBCEllipsoid::infeasibilityStep ( NEWMAT::ColumnVector & ,

NEWMAT::SymmetricMatrix & ,

double &

)

Step taken if the current x is infeasible.

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

Sets up the optimization method.

NLP1* OPTPP::OptBCEllipsoid::nlprob ( ) const [inline, protected]

Pointer to an NLP1*.

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

Call the optimization method.

Implements OPTPP::OptimizeClass.

void OPTPP::OptBCEllipsoid::printStatus ( char * ) [virtual]

Print status of the bound constrained ellipsoidal method.

Implements OPTPP::OptimizeClass.

void OPTPP::OptBCEllipsoid::readOptInput ( ) [virtual]

Read user-specified input options.

Implements OPTPP::OptimizeClass.

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

Implements OPTPP::OptimizeClass.

void OPTPP::OptBCEllipsoid::resetDeepCut ( ) [inline]

Set deepcutflag = 0.

void OPTPP::OptBCEllipsoid::setDeepCut ( ) [inline]

Set deepcutflag = 1.

void OPTPP::OptBCEllipsoid::setInitialEllipsoid ( double rad ) [inline]

Sets the initial ellipsoid radius.

void OPTPP::OptBCEllipsoid::setXScale ( NEWMAT::ColumnVector & x ) [inline]

Set the scaling vector for x.

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

int ndim,

NEWMAT::ColumnVector x

) [inline, virtual]

Implements OPTPP::OptimizeClass.