OPTPP::OptBaNewton Class Reference

#include <OptBaNewton.h>

Inheritance diagram for OPTPP::OptBaNewton:

List of all members.

Public Member Functions
	OptBaNewton (NLP2 *p)
	OptBaNewton (NLP2 *p, UPDATEFCN u)
	OptBaNewton (NLP2 *p, TOLS t)
virtual	~OptBaNewton ()
void	initOpt ()
	Initialize algorithmic parameters.
void	initHessian ()
	Compute the Hessian or its approximation at the initial point.
int	checkInnerConvg (int)
int	checkConvg ()
	Check to see if algorithm satisfies the convergence criterion.
void	acceptStep (int, int)
void	optimize ()
	Invoke a bound constrained Newton's method.
void	updateBarrierMultiplier ()
NEWMAT::SymmetricMatrix	updateH (NEWMAT::SymmetricMatrix &H, int k)
	Compute the Hessian or its approximation at the current point.
double	compute_Barrier_Fvalue (double, NEWMAT::ColumnVector &)
NEWMAT::ColumnVector	compute_Barrier_Gradient (NEWMAT::ColumnVector &, NEWMAT::ColumnVector &)
NEWMAT::SymmetricMatrix	compute_Barrier_Hessian (NEWMAT::SymmetricMatrix &, NEWMAT::ColumnVector &)
NEWMAT::ColumnVector	computeSearch2 (NEWMAT::SymmetricMatrix &, NEWMAT::ColumnVector &)
int	computeStep (NEWMAT::ColumnVector)
	Compute the step length along the search direction.
double	computeMaxStep (NEWMAT::ColumnVector &)
	Compute the maximum step allowed along the search direction before we hit a constraint.
double	scalarNewton (double, double, double, double, double)
void	setAsideCurrentVariables ()
Private Attributes
double	mu
	Barrier parameter.
double	fvalue_barrier
	Value of barrier function.
double	fprev_barrier
	Value of barrier function.
double	fprev_outer
	Value of barrier function.
NEWMAT::ColumnVector	grad_barrier
	Gradient of barrier function.
NEWMAT::ColumnVector	gprev_barrier
	Gradient of barrier function.
NEWMAT::SymmetricMatrix	Hess_barrier
	Hessian of barrier function.

---

Detailed Description

OptBaNewton implements a bound constrained Newton method with a logarithmic barrier term. Barrier methods transform a constrained into an unconstrained problem via a barrier term. In this implementation, the unconstrained subproblem's objective function consists of the original objective function and a logarithmic barrier term that prevents iterates from becoming infeasible.

Author:

J.C. Meza, Lawrence Berkeley National Laboratory

Note:

Modified by P.J. Williams, pwillia@sandia.gov

Date:

Last modified 11/2005

---

Constructor & Destructor Documentation

OPTPP::OptBaNewton::OptBaNewton (  NLP2 *  p  )  [inline]

Parameters: p  a pointer to an NLP2 object. See also: OptNIPSLike(NLP2* p, UPDATEFCN u) OptNIPSLike(NLP2* p, TOLS t)

OPTPP::OptBaNewton::OptBaNewton (  NLP2 *  p, UPDATEFCN  u )  [inline]

Parameters: p  a pointer to an NLP2 object. u  a function pointer. See also: OptNIPSLike(NLP2* p) OptNIPSLike(NLP2* p, TOLS t)

OPTPP::OptBaNewton::OptBaNewton (  NLP2 *  p, TOLS  t )  [inline]

Parameters: p  a pointer to an NLP2 object. t  tolerance class reference. See also: OptNIPSLike(NLP2* p) OptNIPSLike(NLP2* p, UPDATEFCN u)

virtual OPTPP::OptBaNewton::~OptBaNewton (   )  [inline, virtual]

Destructor

---

Member Function Documentation

void OPTPP::OptBaNewton::acceptStep (  int  , int  )  [virtual]

Note: Pure virtual functions Each derived class must define these functions for themselves Reimplemented from OPTPP::OptBCNewtonLike.

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

Check to see if algorithm satisfies the convergence criterion. Reimplemented from OPTPP::OptBCNewtonLike.

int OPTPP::OptBaNewton::checkInnerConvg (  int   )

double OPTPP::OptBaNewton::compute_Barrier_Fvalue (  double  , NEWMAT::ColumnVector &  )

ColumnVector OPTPP::OptBaNewton::compute_Barrier_Gradient (  NEWMAT::ColumnVector &  , NEWMAT::ColumnVector &  )

SymmetricMatrix OPTPP::OptBaNewton::compute_Barrier_Hessian (  NEWMAT::SymmetricMatrix &  , NEWMAT::ColumnVector &  )

double OPTPP::OptBaNewton::computeMaxStep (  NEWMAT::ColumnVector &   )  [virtual]

Compute the maximum step allowed along the search direction before we hit a constraint. Reimplemented from OPTPP::OptBCNewtonLike.

ColumnVector OPTPP::OptBaNewton::computeSearch2 (  NEWMAT::SymmetricMatrix &  , NEWMAT::ColumnVector &  )

int OPTPP::OptBaNewton::computeStep (  NEWMAT::ColumnVector   )  [virtual]

Compute the step length along the search direction. Reimplemented from OPTPP::OptBCNewtonLike.

void OPTPP::OptBaNewton::initHessian (   )  [virtual]

Compute the Hessian or its approximation at the initial point. Reimplemented from OPTPP::OptBCNewtonLike.

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

Initialize algorithmic parameters. Reimplemented from OPTPP::OptBCNewtonLike.

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

Invoke a bound constrained Newton's method. Reimplemented from OPTPP::OptBCNewtonLike.

double OPTPP::OptBaNewton::scalarNewton (  double  , double  , double  , double  , double  )

void OPTPP::OptBaNewton::setAsideCurrentVariables (   )

void OPTPP::OptBaNewton::updateBarrierMultiplier (   )

SymmetricMatrix OPTPP::OptBaNewton::updateH (  NEWMAT::SymmetricMatrix &  H, int  k )  [virtual]

Compute the Hessian or its approximation at the current point. Implements OPTPP::OptBCNewtonLike.

---

Member Data Documentation

double OPTPP::OptBaNewton::fprev_barrier [private]

Value of barrier function.

double OPTPP::OptBaNewton::fprev_outer [private]

Value of barrier function.

double OPTPP::OptBaNewton::fvalue_barrier [private]

Value of barrier function.

NEWMAT::ColumnVector OPTPP::OptBaNewton::gprev_barrier [private]

Gradient of barrier function.

NEWMAT::ColumnVector OPTPP::OptBaNewton::grad_barrier [private]

Gradient of barrier function.

NEWMAT::SymmetricMatrix OPTPP::OptBaNewton::Hess_barrier [private]

Hessian of barrier function.

double OPTPP::OptBaNewton::mu [private]

Barrier parameter.

---

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

• OptBaNewton.h
• OptBaNewton.C

---

Bug Reports    OPT++ Developers    Copyright Information    GNU Lesser General Public License Documentation, generated by , last revised August 30, 2006.