OPTPP::OptConstrNewtonLike Class Reference

Constrained Newton abstract data classes OptConstrNewtonLike OptConstrNewton1Deriv OptConstrNewton2Deriv. More...

Inheritance diagram for OPTPP::OptConstrNewtonLike:
[legend]
Collaboration diagram for OPTPP::OptConstrNewtonLike:
[legend]

List of all members.

Public Member Functions

virtual void acceptStep (int k, int step_type)
 Accept this step and update the nonlinear model.
int checkAnalyticFDGrad ()
 Check finite difference approximation with analytic derivatives.
virtual int checkConvg ()
 Check to see if algorithm satisfies the convergence criterion.
virtual int checkDeriv ()
 Compare the analytic gradient with the finite-difference approximation.
NEWMAT::ColumnVector computeFFK1Ind (const NEWMAT::ColumnVector &xc)
 Compute the active set according to Facchinei, Fischer, and Kanzow indicator.
NEWMAT::ColumnVector computeFFK2Ind (const NEWMAT::ColumnVector &xc)
 Compute the active set according to Facchinei, Fischer, and Kanzow indicator.
virtual real computeMaxStep (NEWMAT::ColumnVector sk)
virtual NEWMAT::ColumnVector computeSearch (NEWMAT::SymmetricMatrix &H)
 Solve for the Newton direction.
virtual int computeStep (NEWMAT::ColumnVector sk)
 Compute the steplength along the search direction.
NEWMAT::ColumnVector computeTapiaInd (const NEWMAT::ColumnVector &step)
 Compute the Tapia indicators for the constrained problem.
void fPrintMultipliers (ostream *nlpout, char *)
 Print the Lagrangian multipliers to a file.
void fPrintSecSuff (ostream *nlpout, NEWMAT::ColumnVector &info)
 Print second order sufficiency information to a file.
NEWMAT::Matrix getConstraintGradient () const
NEWMAT::ColumnVector getConstraintResidual () const
real getCost () const
DerivOption getDerivOption () const
 Set type of finite difference approximation (forward, backward, or central).
NEWMAT::ColumnVector getEqualityMultiplier () const
bool getFeasibilityRecoveryFlag () const
int getFevals () const
int getGevals () const
NEWMAT::ColumnVector getGradL () const
NEWMAT::ColumnVector getGradLPrev () const
real getGradMult () const
NEWMAT::SymmetricMatrix getHessian () const
NEWMAT::ColumnVector getInequalityMultiplier () const
int getMaxFeasIter () const
MeritFcn getMeritFcn () const
int getSearchSize () const
SearchStrategy getSearchStrategy () const
 Set globalization strategy for optimization algorithm.
NEWMAT::ColumnVector getSlacks () const
real getTRSize () const
bool getWarmStart () const
virtual void initHessian ()
 Compute the Hessian or its approximation at the initial point.
virtual void initOpt ()
 Initialize algorithmic parameters.
virtual double initTrustRegionSize () const
 Initialize the radius of the trust-region.
 OptConstrNewtonLike (int n, TOLS t)
 OptConstrNewtonLike (int n, UPDATEFCN u)
 OptConstrNewtonLike (int n)
 OptConstrNewtonLike ()
 Default Constructor.
virtual void optimize ()
 Invoke a constrained Newton's method.
void printMultipliers (char *)
 Output the Lagrangian multipliers to the screen.
void printStatus (char *)
 Print status of the constrained Newton's method.
virtual void readOptInput ()
 Read user-specified input options from a file.
virtual void reset ()
virtual void setConstraintGradient (const NEWMAT::Matrix &constraint_grad)
 Store the current gradients of the constraints.
virtual void setConstraintResidual (const NEWMAT::ColumnVector &constraint_value)
 Store the residuals of the constraints at the current iteration.
void setCost (real value)
 Store current value of merit function.
void setDerivOption (DerivOption d)
void setEqualityMultiplier (const NEWMAT::ColumnVector &ymult)
 Store the current Lagrangian multipliers corresponding to equality constraints.
void setFeasibilityRecoveryFlag (bool flag)
 Set switch to turn on feasibility recovery method.
virtual void setGradL (NEWMAT::ColumnVector gradl_value)
 Store the gradient of the Lagrangian at the current iteration.
virtual void setGradLPrev (NEWMAT::ColumnVector gradl_value)
 Store the gradient of the Lagrangian at the previous iteration.
void setGradMult (real tau)
 Set gradient multiplier which is used to compute radius of trust-region.
void setHessian (NEWMAT::SymmetricMatrix &H)
 Store current Hessian matrix.
void setInequalityMultiplier (const NEWMAT::ColumnVector &zmult)
 Store the current Lagrangian multipliers corresponding to inequality constraints.
void setMaxFeasIter (int k)
 Set maximum number of iterations for feasibility recovery method of trust-region.
virtual void setMeritFcn (MeritFcn option)
 Specify the merit function to used in step acceptance test.
void setSearchSize (int sss)
 Set number of points in search scheme for trustpds strategy.
void setSearchStrategy (SearchStrategy search)
void setSlacks (const NEWMAT::ColumnVector &slackVar)
 Store the current slack vector.
void setTRSize (real delta)
 Set radius of trust-region.
virtual NEWMAT::SymmetricMatrix updateH (NEWMAT::SymmetricMatrix &H, int k)=0
 Compute the Hessian of the Lagrangrian or its approximation at iteration k.
virtual void updateModel (int k, int ndim, NEWMAT::ColumnVector x)
void UseWarmStart (NEWMAT::SymmetricMatrix &H)
virtual ~OptConstrNewtonLike ()
 Destructor.

Protected Member Functions

void defaultAcceptStep (int, int)
 Provide default implementation of AcceptStep.
NEWMAT::ColumnVector defaultComputeSearch (NEWMAT::SymmetricMatrix &)
virtual NLP1 * nlprob () const =0
 returns an NLP1 pointer

Protected Attributes

NEWMAT::Matrix constraintGradient
 Current constraint gradient.
NEWMAT::Matrix constraintGradientPrev
 Previous constraint gradient.
NEWMAT::ColumnVector constraintResidual
 Constraint residual at xc.
NEWMAT::ColumnVector constrType
 Vector which contains the type of constraint.
real cost
 Value of the merit function.
bool feas_flag
 Switch to turn on the feasibility recovery method.
DerivOption finitediff
 User-specified derivative option.
NEWMAT::ColumnVector gprev
 Gradient at previous iteration.
int grad_evals
 Number of gradient evaluations.
NEWMAT::ColumnVector gradl
 Current gradient of the lagrangian.
NEWMAT::ColumnVector gradlprev
 Previous gradient of the lagrangian.
real gradMult
 Gradient multiplier to compute TR_size.
NEWMAT::SymmetricMatrix Hessian
 Current Hessian.
NEWMAT::SymmetricMatrix hessl
 Current Hessian of the lagrangian.
int max_feas_iter
 Maximize number of iterations to achieve feasibility.
int me
 Number of equality constraints.
MeritFcn mfcn
 Merit function option.
int mi
 Number of inequality constraints.
NEWMAT::ColumnVector s
 Slack variables corr. to ineq. constraints.
int searchSize
 Search pattern size for TRPDS.
SearchStrategy strategy
 User-specified globalization strategy.
real TR_size
 Size of the trust region radius.
bool WarmStart
NEWMAT::ColumnVector y
 Lagrange mult. corr. to eq constraints.
NEWMAT::ColumnVector z
 Lagrange mult. corr. to ineq constraints.

Friends

int trustpds (NLP1 *, ostream *, NEWMAT::SymmetricMatrix &, NEWMAT::ColumnVector &, NEWMAT::ColumnVector &, real &, real &, real stpmax, real stpmin, int)
int trustregion (NLP1 *, ostream *, NEWMAT::SymmetricMatrix &, NEWMAT::ColumnVector &, NEWMAT::ColumnVector &, real &, real &, real stpmax, real stpmin)

Detailed Description

Constrained Newton abstract data classes OptConstrNewtonLike OptConstrNewton1Deriv OptConstrNewton2Deriv.

OptConstrNewtonLike provides common data and functionality for OptConstrFDNewton, OptConstrQNewton, OptConstrNewton, OptFDNIPS, OptQNIPS, and OptNIPS methods.

Author:
J.C. Meza, Lawrence Berkeley National Laboratory
Note:
Modified by P.J. Williams
Date:
03/27/2007

Definition at line 31 of file OptConstrNewtonLike.h.


Constructor & Destructor Documentation

OPTPP::OptConstrNewtonLike::OptConstrNewtonLike (  )  [inline]
OPTPP::OptConstrNewtonLike::OptConstrNewtonLike ( int  n  )  [inline]
Parameters:
n an integer argument
See also:
OptConstrNewtonLike(int n, UPDATEFCN u)
OptConstrNewtonLike(int n, TOLS t)

Definition at line 76 of file OptConstrNewtonLike.h.

References s, y, and z.

OPTPP::OptConstrNewtonLike::OptConstrNewtonLike ( int  n,
UPDATEFCN  u 
) [inline]
Parameters:
n an integer argument
u a function pointer.
See also:
OptConstrNewtonLike(int n)
OptConstrNewtonLike(int n, TOLS t)

Definition at line 91 of file OptConstrNewtonLike.h.

References s, OPTPP::OptimizeClass::update_fcn, y, and z.

OPTPP::OptConstrNewtonLike::OptConstrNewtonLike ( int  n,
TOLS  t 
) [inline]
Parameters:
n an integer argument
t tolerance class reference.
See also:
OptConstrNewtonLike(int n)
OptConstrNewtonLike(int n, UPDATEFCN u)

Definition at line 106 of file OptConstrNewtonLike.h.

References s, y, and z.

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

Destructor.

Definition at line 119 of file OptConstrNewtonLike.h.


Member Function Documentation

virtual void OPTPP::OptConstrNewtonLike::acceptStep ( int  k,
int  step_type 
) [inline, virtual]

Accept this step and update the nonlinear model.

Implements OPTPP::OptimizeClass.

Definition at line 127 of file OptConstrNewtonLike.h.

References defaultAcceptStep().

int OPTPP::OptConstrNewtonLike::checkAnalyticFDGrad (  ) 

Check finite difference approximation with analytic derivatives.

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

Check to see if algorithm satisfies the convergence criterion.

Implements OPTPP::OptimizeClass.

Reimplemented in OPTPP::OptNIPSLike.

virtual int OPTPP::OptConstrNewtonLike::checkDeriv (  )  [virtual]

Compare the analytic gradient with the finite-difference approximation.

Reimplemented in OPTPP::OptConstrFDNewton, OPTPP::OptConstrNewton, OPTPP::OptConstrQNewton, OPTPP::OptFDNIPS, OPTPP::OptNIPSLike, and OPTPP::OptQNIPS.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::computeFFK1Ind ( const NEWMAT::ColumnVector &  xc  ) 

Compute the active set according to Facchinei, Fischer, and Kanzow indicator.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::computeFFK2Ind ( const NEWMAT::ColumnVector &  xc  ) 

Compute the active set according to Facchinei, Fischer, and Kanzow indicator.

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

Solve for the Newton direction.

Implements OPTPP::OptimizeClass.

Definition at line 131 of file OptConstrNewtonLike.h.

virtual int OPTPP::OptConstrNewtonLike::computeStep ( NEWMAT::ColumnVector  sk  )  [virtual]

Compute the steplength along the search direction.

If an acceptable step not found, returns an error code = -1.

Reimplemented in OPTPP::OptNIPSLike.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::computeTapiaInd ( const NEWMAT::ColumnVector &  step  ) 

Compute the Tapia indicators for the constrained problem.

void OPTPP::OptConstrNewtonLike::defaultAcceptStep ( int  ,
int   
) [protected, virtual]

Provide default implementation of AcceptStep.

Reimplemented from OPTPP::OptimizeClass.

Referenced by acceptStep().

void OPTPP::OptConstrNewtonLike::fPrintMultipliers ( ostream *  nlpout,
char *   
)

Print the Lagrangian multipliers to a file.

void OPTPP::OptConstrNewtonLike::fPrintSecSuff ( ostream *  nlpout,
NEWMAT::ColumnVector &  info 
)

Print second order sufficiency information to a file.

NEWMAT::Matrix OPTPP::OptConstrNewtonLike::getConstraintGradient (  )  const [inline]
Returns:
Gradient of the constraints at the current iteration

Definition at line 275 of file OptConstrNewtonLike.h.

References constraintGradient.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::getConstraintResidual (  )  const [inline]
Returns:
Residuals of the constraints at the current iteration

Definition at line 267 of file OptConstrNewtonLike.h.

References constraintResidual.

real OPTPP::OptConstrNewtonLike::getCost (  )  const [inline]
Returns:
Value of merit function

Definition at line 283 of file OptConstrNewtonLike.h.

References cost.

DerivOption OPTPP::OptConstrNewtonLike::getDerivOption (  )  const [inline]

Set type of finite difference approximation (forward, backward, or central).

Definition at line 220 of file OptConstrNewtonLike.h.

References finitediff.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::getEqualityMultiplier (  )  const [inline]
Returns:
Lagrangian multipliers corresponding to equality constraints

Definition at line 225 of file OptConstrNewtonLike.h.

References y.

bool OPTPP::OptConstrNewtonLike::getFeasibilityRecoveryFlag (  )  const [inline]
Returns:
Switch to turn on feasibility recovery method

Definition at line 290 of file OptConstrNewtonLike.h.

References feas_flag.

int OPTPP::OptConstrNewtonLike::getFevals (  )  const [inline]
Returns:
The number of function evaluations

Definition at line 176 of file OptConstrNewtonLike.h.

References OPTPP::OptimizeClass::fcn_evals.

int OPTPP::OptConstrNewtonLike::getGevals (  )  const [inline]
Returns:
The number of gradient evaluations

Definition at line 181 of file OptConstrNewtonLike.h.

References grad_evals.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::getGradL (  )  const [inline]
Returns:
Gradient of the Lagrangian at the current iteration

Definition at line 253 of file OptConstrNewtonLike.h.

References gradl.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::getGradLPrev (  )  const [inline]
Returns:
Gradient of the Lagrangian at the previous iteration

Definition at line 260 of file OptConstrNewtonLike.h.

References gradlprev.

real OPTPP::OptConstrNewtonLike::getGradMult (  )  const [inline]
Returns:
Gradient multiplier to compute radius of trust-region

Definition at line 193 of file OptConstrNewtonLike.h.

References gradMult.

NEWMAT::SymmetricMatrix OPTPP::OptConstrNewtonLike::getHessian (  )  const [inline]
Returns:
Hessian matrix

Definition at line 304 of file OptConstrNewtonLike.h.

References Hessian.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::getInequalityMultiplier (  )  const [inline]
Returns:
Lagrangian multipliers corresponding to inequality constraints

Definition at line 232 of file OptConstrNewtonLike.h.

References z.

int OPTPP::OptConstrNewtonLike::getMaxFeasIter (  )  const [inline]
Returns:
Maximum number of iterations for feasibility recovery method

Definition at line 297 of file OptConstrNewtonLike.h.

References max_feas_iter.

MeritFcn OPTPP::OptConstrNewtonLike::getMeritFcn (  )  const [inline]
Returns:
Merit function

Definition at line 246 of file OptConstrNewtonLike.h.

References mfcn.

int OPTPP::OptConstrNewtonLike::getSearchSize (  )  const [inline]
Returns:
Number of points in search scheme (only relevant when trustpds search strategy is selected)

Definition at line 201 of file OptConstrNewtonLike.h.

References searchSize.

SearchStrategy OPTPP::OptConstrNewtonLike::getSearchStrategy (  )  const [inline]

Set globalization strategy for optimization algorithm.

Definition at line 213 of file OptConstrNewtonLike.h.

References strategy.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::getSlacks (  )  const [inline]
Returns:
Slack variables associated with inequality constraints

Definition at line 239 of file OptConstrNewtonLike.h.

References s.

real OPTPP::OptConstrNewtonLike::getTRSize (  )  const [inline]
Returns:
Trust region radius

Definition at line 186 of file OptConstrNewtonLike.h.

References TR_size.

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

Compute the Hessian or its approximation at the initial point.

Reimplemented in OPTPP::OptConstrNewton, OPTPP::OptDHNIPS, OPTPP::OptNIPS, and OPTPP::OptNIPSLike.

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

Initialize algorithmic parameters.

Reimplemented in OPTPP::OptNIPSLike.

virtual double OPTPP::OptConstrNewtonLike::initTrustRegionSize (  )  const [virtual]

Initialize the radius of the trust-region.

virtual NLP1* OPTPP::OptConstrNewtonLike::nlprob (  )  const [protected, pure virtual]
virtual void OPTPP::OptConstrNewtonLike::optimize (  )  [virtual]

Invoke a constrained Newton's method.

Implements OPTPP::OptimizeClass.

Reimplemented in OPTPP::OptNIPSLike.

void OPTPP::OptConstrNewtonLike::printMultipliers ( char *   ) 

Output the Lagrangian multipliers to the screen.

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

Print status of the constrained Newton's method.

Implements OPTPP::OptimizeClass.

Reimplemented in OPTPP::OptConstrNewton, OPTPP::OptDHNIPS, OPTPP::OptNIPS, and OPTPP::OptNIPSLike.

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

Read user-specified input options from a file.

Implements OPTPP::OptimizeClass.

Reimplemented in OPTPP::OptNIPSLike.

virtual void OPTPP::OptConstrNewtonLike::setConstraintGradient ( const NEWMAT::Matrix &  constraint_grad  )  [inline, virtual]

Store the current gradients of the constraints.

Definition at line 277 of file OptConstrNewtonLike.h.

References constraintGradient.

virtual void OPTPP::OptConstrNewtonLike::setConstraintResidual ( const NEWMAT::ColumnVector &  constraint_value  )  [inline, virtual]

Store the residuals of the constraints at the current iteration.

Definition at line 269 of file OptConstrNewtonLike.h.

References constraintResidual.

void OPTPP::OptConstrNewtonLike::setCost ( real  value  )  [inline]

Store current value of merit function.

Definition at line 285 of file OptConstrNewtonLike.h.

References cost.

void OPTPP::OptConstrNewtonLike::setDerivOption ( DerivOption  d  )  [inline]
Returns:
Type of finite difference approximation

Definition at line 218 of file OptConstrNewtonLike.h.

References finitediff.

void OPTPP::OptConstrNewtonLike::setEqualityMultiplier ( const NEWMAT::ColumnVector &  ymult  )  [inline]

Store the current Lagrangian multipliers corresponding to equality constraints.

Definition at line 227 of file OptConstrNewtonLike.h.

References y.

void OPTPP::OptConstrNewtonLike::setFeasibilityRecoveryFlag ( bool  flag  )  [inline]

Set switch to turn on feasibility recovery method.

Definition at line 292 of file OptConstrNewtonLike.h.

References feas_flag.

virtual void OPTPP::OptConstrNewtonLike::setGradL ( NEWMAT::ColumnVector  gradl_value  )  [inline, virtual]

Store the gradient of the Lagrangian at the current iteration.

Definition at line 255 of file OptConstrNewtonLike.h.

References gradl.

virtual void OPTPP::OptConstrNewtonLike::setGradLPrev ( NEWMAT::ColumnVector  gradl_value  )  [inline, virtual]

Store the gradient of the Lagrangian at the previous iteration.

Definition at line 262 of file OptConstrNewtonLike.h.

References gradlprev.

void OPTPP::OptConstrNewtonLike::setGradMult ( real  tau  )  [inline]

Set gradient multiplier which is used to compute radius of trust-region.

Definition at line 195 of file OptConstrNewtonLike.h.

References gradMult.

void OPTPP::OptConstrNewtonLike::setHessian ( NEWMAT::SymmetricMatrix &  H  )  [inline]

Store current Hessian matrix.

Definition at line 306 of file OptConstrNewtonLike.h.

References Hessian.

void OPTPP::OptConstrNewtonLike::setInequalityMultiplier ( const NEWMAT::ColumnVector &  zmult  )  [inline]

Store the current Lagrangian multipliers corresponding to inequality constraints.

Definition at line 234 of file OptConstrNewtonLike.h.

References z.

void OPTPP::OptConstrNewtonLike::setMaxFeasIter ( int  k  )  [inline]

Set maximum number of iterations for feasibility recovery method of trust-region.

Definition at line 299 of file OptConstrNewtonLike.h.

References max_feas_iter.

virtual void OPTPP::OptConstrNewtonLike::setMeritFcn ( MeritFcn  option  )  [inline, virtual]

Specify the merit function to used in step acceptance test.

Reimplemented in OPTPP::OptNIPSLike.

Definition at line 248 of file OptConstrNewtonLike.h.

References mfcn.

void OPTPP::OptConstrNewtonLike::setSearchSize ( int  sss  )  [inline]

Set number of points in search scheme for trustpds strategy.

Definition at line 203 of file OptConstrNewtonLike.h.

References searchSize.

void OPTPP::OptConstrNewtonLike::setSearchStrategy ( SearchStrategy  search  )  [inline]
Returns:
Globalization strategy for optimization algorithm

Definition at line 211 of file OptConstrNewtonLike.h.

References strategy.

void OPTPP::OptConstrNewtonLike::setSlacks ( const NEWMAT::ColumnVector &  slackVar  )  [inline]

Store the current slack vector.

Definition at line 241 of file OptConstrNewtonLike.h.

References s.

void OPTPP::OptConstrNewtonLike::setTRSize ( real  delta  )  [inline]

Set radius of trust-region.

Definition at line 188 of file OptConstrNewtonLike.h.

References TR_size.

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

Compute the Hessian of the Lagrangrian or its approximation at iteration k.

Implemented in OPTPP::OptConstrFDNewton, OPTPP::OptConstrNewton, OPTPP::OptConstrQNewton, OPTPP::OptDHNIPS, OPTPP::OptFDNIPS, OPTPP::OptNIPS, OPTPP::OptNIPSLike, and OPTPP::OptQNIPS.


Member Data Documentation

Current constraint gradient.

Definition at line 45 of file OptConstrNewtonLike.h.

Referenced by getConstraintGradient(), and setConstraintGradient().

Previous constraint gradient.

Definition at line 46 of file OptConstrNewtonLike.h.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::constraintResidual [protected]

Constraint residual at xc.

Definition at line 42 of file OptConstrNewtonLike.h.

Referenced by getConstraintResidual(), and setConstraintResidual().

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::constrType [protected]

Vector which contains the type of constraint.

Definition at line 41 of file OptConstrNewtonLike.h.

Value of the merit function.

Definition at line 55 of file OptConstrNewtonLike.h.

Referenced by getCost(), and setCost().

Switch to turn on the feasibility recovery method.

Definition at line 59 of file OptConstrNewtonLike.h.

Referenced by getFeasibilityRecoveryFlag(), and setFeasibilityRecoveryFlag().

DerivOption OPTPP::OptConstrNewtonLike::finitediff [protected]

User-specified derivative option.

Definition at line 50 of file OptConstrNewtonLike.h.

Referenced by getDerivOption(), and setDerivOption().

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::gprev [protected]

Gradient at previous iteration.

Definition at line 37 of file OptConstrNewtonLike.h.

Number of gradient evaluations.

Definition at line 36 of file OptConstrNewtonLike.h.

Referenced by getGevals().

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::gradl [protected]

Current gradient of the lagrangian.

Definition at line 43 of file OptConstrNewtonLike.h.

Referenced by getGradL(), and setGradL().

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::gradlprev [protected]

Previous gradient of the lagrangian.

Definition at line 44 of file OptConstrNewtonLike.h.

Referenced by getGradLPrev(), and setGradLPrev().

Gradient multiplier to compute TR_size.

Definition at line 53 of file OptConstrNewtonLike.h.

Referenced by getGradMult(), and setGradMult().

NEWMAT::SymmetricMatrix OPTPP::OptConstrNewtonLike::Hessian [protected]

Current Hessian.

Definition at line 47 of file OptConstrNewtonLike.h.

Referenced by getHessian(), and setHessian().

NEWMAT::SymmetricMatrix OPTPP::OptConstrNewtonLike::hessl [protected]

Current Hessian of the lagrangian.

Definition at line 48 of file OptConstrNewtonLike.h.

Maximize number of iterations to achieve feasibility.

Definition at line 60 of file OptConstrNewtonLike.h.

Referenced by getMaxFeasIter(), and setMaxFeasIter().

Number of equality constraints.

Definition at line 34 of file OptConstrNewtonLike.h.

MeritFcn OPTPP::OptConstrNewtonLike::mfcn [protected]

Merit function option.

Definition at line 51 of file OptConstrNewtonLike.h.

Referenced by getMeritFcn(), and setMeritFcn().

Number of inequality constraints.

Definition at line 35 of file OptConstrNewtonLike.h.

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::s [protected]

Slack variables corr. to ineq. constraints.

Definition at line 40 of file OptConstrNewtonLike.h.

Referenced by getSlacks(), OptConstrNewtonLike(), and setSlacks().

Search pattern size for TRPDS.

Definition at line 54 of file OptConstrNewtonLike.h.

Referenced by getSearchSize(), and setSearchSize().

SearchStrategy OPTPP::OptConstrNewtonLike::strategy [protected]

User-specified globalization strategy.

Definition at line 49 of file OptConstrNewtonLike.h.

Referenced by getSearchStrategy(), and setSearchStrategy().

Size of the trust region radius.

Definition at line 52 of file OptConstrNewtonLike.h.

Referenced by getTRSize(), and setTRSize().

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::y [protected]

Lagrange mult. corr. to eq constraints.

Definition at line 39 of file OptConstrNewtonLike.h.

Referenced by getEqualityMultiplier(), OptConstrNewtonLike(), and setEqualityMultiplier().

NEWMAT::ColumnVector OPTPP::OptConstrNewtonLike::z [protected]

Lagrange mult. corr. to ineq constraints.

Definition at line 38 of file OptConstrNewtonLike.h.

Referenced by getInequalityMultiplier(), OptConstrNewtonLike(), and setInequalityMultiplier().


The documentation for this class was generated from the following file:
Generated on Mon Jan 24 12:04:41 2011 for FASTlib by  doxygen 1.6.3