Linear least squares. More...

#include <vnl_lsqr.h>

Public Member Functions
	vnl_lsqr (vnl_linear_system &ls)
	~vnl_lsqr ()
void	set_max_iterations (long max_iter)
int	minimize (vnl_vector< double > &x)
	Perform the minimization starting from x=0 and putting the result into x.
long	get_number_of_iterations () const
void	diagnose_outcome (vcl_ostream &os) const
	Pontificate about the outcome of the last minimization.
double	get_resid_norm_estimate () const
	Return the residual norm estimate:.
int	return_code () const
	Get the return code for the last minimization.
Static Public Member Functions
static void	translate_return_code (vcl_ostream &os, int return_code)
Static Protected Member Functions
static int	aprod_ (long mode, long m, long n, double x, double y, long leniw, long lenrw, long iw, double rw, void userdata)
Protected Attributes
vnl_linear_system *	ls_
long	max_iter_
long	num_iter_
double	resid_norm_estimate_
double	result_norm_estimate_
double	A_condition_estimate_
double	result_norm_
long	return_code_

Detailed Description

Linear least squares.

vnl_lsqr implements an algorithm for large, sparse linear systems and sparse, linear least squares. It is a wrapper for the LSQR algorithm of Paige and Saunders (ACM TOMS 583). The sparse system is encapsulated by a vnl_linear_system.

Definition at line 35 of file vnl_lsqr.h.

Constructor & Destructor Documentation

vnl_lsqr::vnl_lsqr ( vnl_linear_system & ls ) [inline]

Definition at line 38 of file vnl_lsqr.h.

vnl_lsqr::~vnl_lsqr ( )

Definition at line 19 of file vnl_lsqr.cxx.

Member Function Documentation

int vnl_lsqr::aprod_	(	long *	mode,
		long *	m,
		long *	n,
		double *	x,
		double *	y,
		long *	leniw,
		long *	lenrw,
		long *	iw,
		double *	rw,
		void *	userdata
	)		`[static, protected]`