Parameters to control the minimization algorithm. More...

#include <MinimizeParams.h>

Public Types
enum	DirectionUpdateScheme { PolakRibiere, FletcherReeves, HestenesStiefel, LBFGS, SteepestDescent }
	Search direction update scheme. More...

enum	LinminMethod { DirUpdateRecommended, Relax, Quad, CubicWolfe }
	Line minimization method. More...

Public Member Functions
	MinimizeParams ()
	Set the default values.

Public Attributes
enum MinimizeParams::DirectionUpdateScheme	dirUpdateScheme

enum MinimizeParams::LinminMethod	linminMethod

int	nIterations
	Maximum number of iterations (default 100)

int	nDim
	Dimension of optimization space; used only for knormThreshold (default 1)

int	history
	Number of past variables and residuals to store (BFGS only)

FILE *	fpLog
	Stream to log iterations to.

const char *	linePrefix
	prefix for each output line of minimizer, useful for nested minimizations (default "CG\t")

const char *	energyLabel
	Label for the minimized quantity (default "E")

const char *	energyFormat
	printf format for the minimized quantity (default "%22.15le")

double	knormThreshold
	stop when norm of residual against preconditioner falls below this (default: 0)

double	energyDiffThreshold
	stop when energy change is below this for nEnergyDiff successive iterations (default: 0)

int	nEnergyDiff
	number of successive iterations for energyDiffThreshold check (default: 2)

double	alphaTstart
	initial value for the test-step size (default: 1.0)

double	alphaTmin
	minimum value of the test-step size (algorithm gives up when difficulties cause alphaT to fall below this value) (default:1e-10)

bool	updateTestStepSize
	set alphaT=alpha after every iteration if true (default: true)

double	alphaTreduceFactor
	Factor to reduce alphaT on difficulties (default 0.1)

double	alphaTincreaseFactor
	Max ratio of alpha to alphaT, increase alphaT by this factor otherwise (default 3.0)

int	nAlphaAdjustMax
	maximum number of times to alpha adjust attempts (default 3)

double	wolfeEnergy
	Wolfe criterion dimensionless threshold for energy.

double	wolfeGradient
	Wolfe criterion dimensionless threshold for gradient.

bool	fdTest
	whether to perform a finite difference test before each minimization (default false)

Detailed Description

Parameters to control the minimization algorithm.

Member Enumeration Documentation

enum MinimizeParams::DirectionUpdateScheme

Search direction update scheme.

Enumerator
PolakRibiere	Polak-Ribiere (preconditioned) conjugate gradients (default)
FletcherReeves	Fletcher-Reeves (preconditioned) conjugate gradients.
HestenesStiefel	Hestenes-Stiefel (preconditioned) conjugate gradients.
LBFGS	Limited memory version of the BFGS algorithm.
SteepestDescent	Steepest Descent (always along negative (preconditioned) gradient)

enum MinimizeParams::LinminMethod

Line minimization method.

Enumerator
DirUpdateRecommended	default method recommended by the direction update scheme
Relax	move by a constant multiple (=alphaTstart) of the search direction (not recommended for CG)
Quad	use the energy at a test step location to find the minimum along the line (default)
CubicWolfe	Cubic line search terminated by Wolfe conditions, possibly without a test step.

The documentation for this struct was generated from the following file:

core/MinimizeParams.h

Public Types

Public Member Functions

Public Attributes

Detailed Description

Member Enumeration Documentation