GClasses

GClasses::GTargetFunction Class Reference

The optimizer seeks to find values that minimize this target function. More...

#include <GOptimizer.h>

List of all members.

Public Member Functions

 GTargetFunction (sp_relation &pRelation)
 GTargetFunction (size_t dims)
virtual ~GTargetFunction ()
sp_relationrelation ()
 Returns a (smart) pointer to the relation, which specifies the type (discrete or real) of each element in the vector that is being optimized.
virtual bool isStable ()=0
 Return true if computeError is completely deterministic with respect to the vector being optimized. Return false if the error also depends on some state other than the vector being optimized. This mostly affects whether the optimization algorithms are permitted to remember old error values for efficiency purposes.
virtual bool isConstrained ()=0
 Return true if this function is constrained to only support certain vectors.
virtual void initVector (double *pVector)=0
 Sets pVector to an initial guess.
virtual double computeError (const double *pVector)=0
 Computes the error of the given vector using all patterns.
virtual double computeErrorOnline (const double *pVector, size_t nPattern)
 Estimates the error of the given vector using a single (usually randomly selected) pattern.
virtual void constrain (double *pVector)
 Adjust pVector to the nearest vector that fits the constraints.

Protected Attributes

sp_relation m_pRelation

Detailed Description

The optimizer seeks to find values that minimize this target function.


Constructor & Destructor Documentation

GClasses::GTargetFunction::GTargetFunction ( sp_relation pRelation) [inline]
GClasses::GTargetFunction::GTargetFunction ( size_t  dims)
virtual GClasses::GTargetFunction::~GTargetFunction ( ) [inline, virtual]

Member Function Documentation

virtual double GClasses::GTargetFunction::computeError ( const double *  pVector) [pure virtual]

Computes the error of the given vector using all patterns.

virtual double GClasses::GTargetFunction::computeErrorOnline ( const double *  pVector,
size_t  nPattern 
) [inline, virtual]

Estimates the error of the given vector using a single (usually randomly selected) pattern.

virtual void GClasses::GTargetFunction::constrain ( double *  pVector) [inline, virtual]

Adjust pVector to the nearest vector that fits the constraints.

virtual void GClasses::GTargetFunction::initVector ( double *  pVector) [pure virtual]

Sets pVector to an initial guess.

virtual bool GClasses::GTargetFunction::isConstrained ( ) [pure virtual]

Return true if this function is constrained to only support certain vectors.

virtual bool GClasses::GTargetFunction::isStable ( ) [pure virtual]

Return true if computeError is completely deterministic with respect to the vector being optimized. Return false if the error also depends on some state other than the vector being optimized. This mostly affects whether the optimization algorithms are permitted to remember old error values for efficiency purposes.

sp_relation& GClasses::GTargetFunction::relation ( ) [inline]

Returns a (smart) pointer to the relation, which specifies the type (discrete or real) of each element in the vector that is being optimized.


Member Data Documentation