Detailed Description

Options for Wolfe line search during optimization.

These parameters control the behavior of the line search, including the Wolfe condition thresholds, step size bounds, and numerical tolerances.

Definition at line 29 of file wolfe_line_search.hpp.

Public Member Functions
constexpr	laplace_line_search_options (int max_iter)

constexpr	laplace_line_search_options ()=default

Public Attributes
int	max_iterations {1000}

double	c1 {1e-4}
	Armijo condition parameter (sufficient decrease).

double	c2 {0.9}
	Curvature condition parameter.

double	tau {0.5}
	Backtracking shrinkage factor.

double	min_alpha {1e-8}
	Minimum allowable step size.

double	max_alpha {16.0}
	Maximum allowable step size.

double	scale_up {2.0}
	Step size expansion factor.

double	abs_grad_threshold {1e-12}
	Absolute gradient tolerance.

double	abs_obj_threshold {1e-12}
	Absolute objective tolerance.

double	rel_grad_threshold {1e-3}

double	rel_obj_threshold {1e-10}

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

Public Member Functions