Class ConfigGaussNewton

java.lang.Object

org.ddogleg.optimization.ConfigGaussNewton

Direct Known Subclasses:

ConfigLevenbergMarquardt, ConfigTrustRegion

public class ConfigGaussNewton
extends Object

Configuration for GaussNewtonBase_F64.

Field Summary

Fields

Modifier and Type	Field	Description
double	ftol	tolerance for termination, change in function value.
double	gtol	tolerance for termination.
boolean	hessianScaling	Optional scaling of Jacobian to make the Hessian matrix better suited for decomposition by improving the matrice's condition scaling = sqrt(diag(B)) where B is the Hessian matrix.
LinearSolverType	solverType	Which linear solver should it use internally.

Constructor Summary

Constructors

Constructor	Description
ConfigGaussNewton()

Method Summary

Modifier and Type	Method	Description
void	reset()
ConfigGaussNewton	setTo(ConfigGaussNewton config)

Methods inherited from class Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

gtol

public double gtol

tolerance for termination. magnitude of gradient. absolute

ftol

public double ftol

tolerance for termination, change in function value. relative

hessianScaling

public boolean hessianScaling

Optional scaling of Jacobian to make the Hessian matrix better suited for decomposition by improving the matrice's condition scaling = sqrt(diag(B)) where B is the Hessian matrix. For Least-Squares B = J'*J

solverType

public LinearSolverType solverType

Which linear solver should it use internally. Default will be what most people will use, but other choices could be faster or more stable. If the specific implementation does not use a linear solver this must be default.

Constructor Details

ConfigGaussNewton

public ConfigGaussNewton()

Method Details

setTo

public ConfigGaussNewton setTo(ConfigGaussNewton config)

reset

public void reset()