Solver options for small-dimension Newton methods. More...

#include <MathSys_NewtonTypes.hxx>

Inheritance diagram for MathSys::NewtonOptions:

Public Member Functions
	NewtonOptions ()
	Default constructor with strict residual/step tolerances for specialized Newton.

Public Member Functions inherited from MathUtils::Config
	Config ()=default
	Default constructor with standard tolerances.

	Config (double theTolerance, int theMaxIter=100)
	Constructor with custom tolerance (sets all tolerances to same value).

Data Fields
double	MaxStepRatio = 0.5
	Max step as ratio of largest domain size.

bool	EnableLineSearch = true
	Enable Armijo backtracking line search.

bool	AllowSoftBounds = false
	Allow slight bounds extension.

double	SoftBoundsExtension = 1.0e-4
	Extension ratio for soft bounds.

Data Fields inherited from MathUtils::Config
int	MaxIterations = 100
	Maximum number of iterations allowed.

double	Tolerance = 1.0e-10
	General convergence tolerance.

double	XTolerance = 1.0e-10
	Tolerance for solution change \|x_{n+1} - x_n\|.

double	FTolerance = 1.0e-10
	Tolerance for function value \|f(x)\|.

double	StepMin = std::numeric_limits<double>::epsilon()
	Minimum step size before declaring convergence or failure.

Detailed Description

Solver options for small-dimension Newton methods.

Constructor & Destructor Documentation

MathSys::NewtonOptions::NewtonOptions ( )

inline

Default constructor with strict residual/step tolerances for specialized Newton.

bool MathSys::NewtonOptions::AllowSoftBounds = false

Allow slight bounds extension.

bool MathSys::NewtonOptions::EnableLineSearch = true

Enable Armijo backtracking line search.

double MathSys::NewtonOptions::MaxStepRatio = 0.5

Max step as ratio of largest domain size.

double MathSys::NewtonOptions::SoftBoundsExtension = 1.0e-4

Extension ratio for soft bounds.

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