---

RazorBack::Lu_ Class Reference

Lu_: a class for doing LU-decomposition. More...

#include <Lu.hh>

List of all members.

Public Methods
	Lu_ (unsigned int N=3)
	Inits to an N x N problem (default 3).
	Lu_ (const Rmatdbl_ &A)
	Inits with the decomposition of the square matrix A.
int	decomp (const Rmatdbl_ &A)
	decomp(A): performs an LU-decomposition on the square matrix A. More...
double	det () const
	det(): calculates the determinant from the decomposition results stored inside. More...
Vecdbl_	solve (const Vecdbl_ &B) const
	solve(B): solves the linear equation A*x=B and returns x. More...
int	lineq (const Rmatdbl_ &A, const Vecdbl_ &B, Vecdbl_ &X, unsigned int Maxit=0)
	lineq(A, B, X, Maxit): solves the nxn linear equation A*X=B and returns X. More...
Private Attributes
Rmatdbl_	Lu
	holds the decomposition.
vector<unsigned int>	Perm
	permutation vector.
int	Psign
	permutation sign.
Static Private Attributes
const double	LU_EPSILON
	a very small number.

---

Detailed Description

Lu_: a class for doing LU-decomposition.

Can be asked to calculate determinants and solve linear equation systems.

Definition at line 33 of file Lu.hh.

---

Constructor & Destructor Documentation

RazorBack::Lu_::Lu_ ( unsigned int N = 3 ) [inline, explicit]

Inits to an N x N problem (default 3). Definition at line 48 of file Lu.hh.

RazorBack::Lu_::Lu_ ( const Rmatdbl_ & A ) [inline, explicit]

Inits with the decomposition of the square matrix A. Definition at line 51 of file Lu.hh.

---

Member Function Documentation

int RazorBack::Lu_::decomp ( const Rmatdbl_ & A )

decomp(A): performs an LU-decomposition on the square matrix A. Return value: the permutation sign (0 if A was singular). Throws Squarexc_ if A was not NxN. Referenced by Lu_().

double RazorBack::Lu_::det ( ) const

det(): calculates the determinant from the decomposition results stored inside. It is assumed that decomp() has been called beforehand.

Vecdbl_ RazorBack::Lu_::solve ( const Vecdbl_ & B ) const

solve(B): solves the linear equation A*x=B and returns x. The calling object is supposed to have been primed with decomp(A) before this call. B is the "right-hand-side" vector which is preserved. Throws a Dimexc_ exception if B's dimension does not match A's.

int RazorBack::Lu_::lineq ( const Rmatdbl_ & A, const Vecdbl_ & B, Vecdbl_ & X, unsigned int Maxit = 0 )

lineq(A, B, X, Maxit): solves the nxn linear equation A*X=B and returns X. Iterative improvement is done a la Num. Recipes (max. Maxit times, default 0). This routine should be used if only one right-hand side vector B is present. For several right-hand sides, use the decomposition and solver routines separately. Return value: 1 means OK, 0 means A was singular. Throws Squarexc_ if A was not NxN, Dimexc_ on dim mismatches.

---

Member Data Documentation

const double RazorBack::Lu_::LU_EPSILON [static, private]

a very small number. Definition at line 38 of file Lu.hh.

Rmatdbl_ RazorBack::Lu_::Lu [private]

holds the decomposition. Definition at line 40 of file Lu.hh.

vector< unsigned int > RazorBack::Lu_::Perm<unsigned int> [private]

permutation vector. Definition at line 41 of file Lu.hh.

int RazorBack::Lu_::Psign [private]

permutation sign. Definition at line 42 of file Lu.hh.

---

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

• Lu.hh

---