Detailed Description

Allows setting and getting of options for OptimizeBFGS instance via IterativeSolver base class.

Public Member Functions
	OptimizeBFGSOptions ()=default

	OptimizeBFGSOptions (const options_map &opt)

Public Member Functions inherited from molpro::linalg::itsolv::OptimizeOptions
	OptimizeOptions ()=default

	OptimizeOptions (const options_map &opt)

Public Member Functions inherited from molpro::linalg::itsolv::Options
virtual	~Options ()=default

	Options ()=default

	Options (const options_map &opt)
	Initialises options from key/value strings. Unrecognised options are ignored. More...

void	copy (const Options &source)
	copies options from source object More...

Public Attributes
std::optional< int >	max_size_qspace

std::optional< double >	norm_thresh

std::optional< double >	svd_thresh

std::optional< bool >	strong_Wolfe
	Whether to use strong or weak Wolfe conditions. More...

std::optional< double >	Wolfe_1
	Acceptance parameter for function value. More...

std::optional< double >	Wolfe_2
	Acceptance parameter for function gradient. More...

std::optional< double >	linesearch_tolerance
	If the predicted line search is within tolerance of 1, don't bother taking it. More...

std::optional< double >	linesearch_grow_factor

std::optional< double >	quasinewton_maximum_step
	Quasi-Newton steps with an L-2 norm larger than this will be scaled down to this value. More...

Public Attributes inherited from molpro::linalg::itsolv::Options
std::optional< double >	convergence_threshold
	convergence threshold More...

std::optional< int >	n_roots
	number of roots to solve for More...

std::optional< Verbosity >	verbosity
	verbosity level during IterativeSolver::solve() More...

std::optional< int >	max_iter
	maximum number of iterations during IterativeSolver::solve() More...

std::optional< double >	max_p
	maximum size of P space More...

std::optional< double >	p_threshold
	threshold for selecting P space More...

Constructor & Destructor Documentation

◆ OptimizeBFGSOptions() [1/2]

molpro::linalg::itsolv::OptimizeBFGSOptions::OptimizeBFGSOptions ( )

default

◆ OptimizeBFGSOptions() [2/2]

molpro::linalg::itsolv::OptimizeBFGSOptions::OptimizeBFGSOptions ( const options_map & opt )

Member Data Documentation

◆ linesearch_grow_factor

std::optional<double> molpro::linalg::itsolv::OptimizeBFGSOptions::linesearch_grow_factor

If the predicted line search step is extrapolation, limit the step to this factor times the current step

◆ linesearch_tolerance

std::optional<double> molpro::linalg::itsolv::OptimizeBFGSOptions::linesearch_tolerance

If the predicted line search is within tolerance of 1, don't bother taking it.

◆ max_size_qspace

std::optional<int> molpro::linalg::itsolv::OptimizeBFGSOptions::max_size_qspace

◆ norm_thresh

std::optional<double> molpro::linalg::itsolv::OptimizeBFGSOptions::norm_thresh

◆ quasinewton_maximum_step

std::optional<double> molpro::linalg::itsolv::OptimizeBFGSOptions::quasinewton_maximum_step

Quasi-Newton steps with an L-2 norm larger than this will be scaled down to this value.

◆ strong_Wolfe

std::optional<bool> molpro::linalg::itsolv::OptimizeBFGSOptions::strong_Wolfe

Whether to use strong or weak Wolfe conditions.

◆ svd_thresh

std::optional<double> molpro::linalg::itsolv::OptimizeBFGSOptions::svd_thresh

◆ Wolfe_1

std::optional<double> molpro::linalg::itsolv::OptimizeBFGSOptions::Wolfe_1

Acceptance parameter for function value.

◆ Wolfe_2

std::optional<double> molpro::linalg::itsolv::OptimizeBFGSOptions::Wolfe_2

Acceptance parameter for function gradient.