GClasses::GAnnealing Class Reference

This algorithm tries the current direction and a slightly perturbed direction at each step. If the perturbed direction resulted in faster improvement, it becomes the new current direction. As long as the current direction yields improvement, it accelerates, otherwise it decelerates. More...

#include <GHillClimber.h>

Inheritance diagram for GClasses::GAnnealing:

Public Member Functions
	GAnnealing (GTargetFunction *pTargetFunc, double initialDeviation, double decay, GRand *pRand)
virtual	~GAnnealing ()
virtual double *	currentVector ()
	Returns the best vector yet found. More...
virtual double	iterate ()
	Performs a little more optimization. (Call this in a loop until acceptable results are found.) More...
void	setDeviation (double d)
	Specify the current deviation to use for annealing. (A random vector from a Normal distribution with the specified deviation will be added to each candidate vector in order to simulate annealing.) More...
Public Member Functions inherited from GClasses::GOptimizer
	GOptimizer (GTargetFunction *pCritic)
virtual	~GOptimizer ()
double	searchUntil (size_t nBurnInIterations, size_t nIterations, double dImprovement)
	This will first call iterate() nBurnInIterations times, then it will repeatedly call iterate() in blocks of nIterations times. If the error heuristic has not improved by the specified ratio after a block of iterations, it will stop. (For example, if the error before the block of iterations was 50, and the error after is 49, then training will stop if dImprovement is > 0.02.) If the error heuristic is not stable, then the value of nIterations should be large. More...

Protected Member Functions
void	reset ()

Protected Attributes
double	m_decay
double	m_dError
double	m_deviation
size_t	m_dims
double	m_initialDeviation
double *	m_pBuf
double *	m_pCandidate
GRand *	m_pRand
double *	m_pVector
Protected Attributes inherited from GClasses::GOptimizer
GTargetFunction *	m_pCritic

Detailed Description

This algorithm tries the current direction and a slightly perturbed direction at each step. If the perturbed direction resulted in faster improvement, it becomes the new current direction. As long as the current direction yields improvement, it accelerates, otherwise it decelerates.

Constructor & Destructor Documentation

GClasses::GAnnealing::GAnnealing	(	GTargetFunction *	pTargetFunc,
		double	initialDeviation,
		double	decay,
		GRand *	pRand
	)

virtual GClasses::GAnnealing::~GAnnealing ( )

virtual

Member Function Documentation

virtual double* GClasses::GAnnealing::currentVector ( )

inlinevirtual

Returns the best vector yet found.

Implements GClasses::GOptimizer.

virtual double GClasses::GAnnealing::iterate ( )

virtual

Performs a little more optimization. (Call this in a loop until acceptable results are found.)

Implements GClasses::GOptimizer.

void GClasses::GAnnealing::reset ( )

protected

void GClasses::GAnnealing::setDeviation ( double  d )

inline

Specify the current deviation to use for annealing. (A random vector from a Normal distribution with the specified deviation will be added to each candidate vector in order to simulate annealing.)

Member Data Documentation

double GClasses::GAnnealing::m_decay

protected

double GClasses::GAnnealing::m_dError

protected

double GClasses::GAnnealing::m_deviation

protected

size_t GClasses::GAnnealing::m_dims

protected

double GClasses::GAnnealing::m_initialDeviation

protected

double* GClasses::GAnnealing::m_pBuf

protected

double* GClasses::GAnnealing::m_pCandidate

protected

GRand* GClasses::GAnnealing::m_pRand

protected

double* GClasses::GAnnealing::m_pVector

protected