This class implements LSQR. More...
#include <ClpLsqr.hpp>
Collaboration diagram for ClpLsqr:
Public Member Functions | |
Constructors and destructors | |
| ClpLsqr () | |
| Default constructor. More... | |
| ClpLsqr (ClpInterior *model) | |
| Constructor for use with Pdco model (note modified for pdco!!!!) More... | |
| ClpLsqr (const ClpLsqr &) | |
| Copy constructor. More... | |
| ClpLsqr & | operator= (const ClpLsqr &rhs) |
| Assignment operator. This copies the data. More... | |
| ~ClpLsqr () | |
| Destructor. More... | |
Methods | |
| bool | setParam (char *parmName, int parmValue) |
| Set an int parameter. More... | |
| void | do_lsqr (CoinDenseVector< double > &b, double damp, double atol, double btol, double conlim, int itnlim, bool show, Info info, CoinDenseVector< double > &x, int *istop, int *itn, Outfo *outfo, bool precon, CoinDenseVector< double > &Pr) |
| Call the Lsqr algorithm. More... | |
| void | matVecMult (int, CoinDenseVector< double > *, CoinDenseVector< double > *) |
| Matrix-vector multiply - implemented by user. More... | |
| void | matVecMult (int, CoinDenseVector< double > &, CoinDenseVector< double > &) |
| Set an int parameter. More... | |
| void | borrowDiag1 (double *array) |
| diag1 - we just borrow as it is part of a CoinDenseVector<double> More... | |
Public Attributes | |
Public member data | |
| int | nrows_ |
| Row dimension of matrix. More... | |
| int | ncols_ |
| Column dimension of matrix. More... | |
| ClpInterior * | model_ |
| Pointer to Model object for this instance. More... | |
| double * | diag1_ |
| Diagonal array 1. More... | |
| double | diag2_ |
| Constant diagonal 2. More... | |
Detailed Description
This class implements LSQR.
LSQR solves Ax = b or min ||b - Ax||_2 if damp = 0,
or min || (b) - ( A )x || otherwise.
|| (0) (damp I) ||2
A is an m by n matrix defined by user provided routines
matVecMult(mode, y, x)
which performs the matrix-vector operations where y and x
are references or pointers to CoinDenseVector objects.
If mode = 1, matVecMult must return y = Ax without altering x.
If mode = 2, matVecMult must return y = A'x without altering x.
-----------------------------------------------------------------------
LSQR uses an iterative (conjugate-gradient-like) method.
For further information, see
1. C. C. Paige and M. A. Saunders (1982a).
LSQR: An algorithm for sparse linear equations and sparse least squares,
ACM TOMS 8(1), 43-71.
2. C. C. Paige and M. A. Saunders (1982b).
Algorithm 583. LSQR: Sparse linear equations and least squares problems,
ACM TOMS 8(2), 195-209.
3. M. A. Saunders (1995). Solution of sparse rectangular systems using
LSQR and CRAIG, BIT 35, 588-604.
Input parameters:
atol, btol are stopping tolerances. If both are 1.0e-9 (say),
the final residual norm should be accurate to about 9 digits.
(The final x will usually have fewer correct digits,
depending on cond(A) and the size of damp.)
conlim is also a stopping tolerance. lsqr terminates if an estimate
of cond(A) exceeds conlim. For compatible systems Ax = b,
conlim could be as large as 1.0e+12 (say). For least-squares
problems, conlim should be less than 1.0e+8.
Maximum precision can be obtained by setting
atol = btol = conlim = zero, but the number of iterations
may then be excessive.
itnlim is an explicit limit on iterations (for safety).
show = 1 gives an iteration log,
show = 0 suppresses output.
info is a structure special to pdco.m, used to test if
was small enough, and continuing if necessary with smaller atol.
Output parameters:
x is the final solution.
*istop gives the reason for termination.
*istop = 1 means x is an approximate solution to Ax = b.
= 2 means x approximately solves the least-squares problem.
rnorm = norm(r) if damp = 0, where r = b - Ax,
= sqrt( norm(r)**2 + damp**2 * norm(x)**2 ) otherwise.
xnorm = norm(x).
var estimates diag( inv(A'A) ). Omitted in this special version.
outfo is a structure special to pdco.m, returning information
about whether atol had to be reduced.
Other potential output parameters:
anorm, acond, arnorm, xnorm
Definition at line 76 of file ClpLsqr.hpp.
Constructor & Destructor Documentation
| ClpLsqr::ClpLsqr | ( | ) |
Default constructor.
| ClpLsqr::ClpLsqr | ( | ClpInterior * | model | ) |
Constructor for use with Pdco model (note modified for pdco!!!!)
| ClpLsqr::ClpLsqr | ( | const ClpLsqr & | ) |
Copy constructor.
| ClpLsqr::~ClpLsqr | ( | ) |
Destructor.
Member Function Documentation
| bool ClpLsqr::setParam | ( | char * | parmName, |
| int | parmValue | ||
| ) |
Set an int parameter.
| void ClpLsqr::do_lsqr | ( | CoinDenseVector< double > & | b, |
| double | damp, | ||
| double | atol, | ||
| double | btol, | ||
| double | conlim, | ||
| int | itnlim, | ||
| bool | show, | ||
| Info | info, | ||
| CoinDenseVector< double > & | x, | ||
| int * | istop, | ||
| int * | itn, | ||
| Outfo * | outfo, | ||
| bool | precon, | ||
| CoinDenseVector< double > & | Pr | ||
| ) |
Call the Lsqr algorithm.
| void ClpLsqr::matVecMult | ( | int | , |
| CoinDenseVector< double > * | , | ||
| CoinDenseVector< double > * | |||
| ) |
Matrix-vector multiply - implemented by user.
| void ClpLsqr::matVecMult | ( | int | , |
| CoinDenseVector< double > & | , | ||
| CoinDenseVector< double > & | |||
| ) |
Set an int parameter.
|
inline |
diag1 - we just borrow as it is part of a CoinDenseVector<double>
Definition at line 125 of file ClpLsqr.hpp.
Member Data Documentation
| int ClpLsqr::nrows_ |
Row dimension of matrix.
Definition at line 86 of file ClpLsqr.hpp.
| int ClpLsqr::ncols_ |
Column dimension of matrix.
Definition at line 88 of file ClpLsqr.hpp.
| ClpInterior* ClpLsqr::model_ |
Pointer to Model object for this instance.
Definition at line 90 of file ClpLsqr.hpp.
| double* ClpLsqr::diag1_ |
Diagonal array 1.
Definition at line 92 of file ClpLsqr.hpp.
| double ClpLsqr::diag2_ |
Constant diagonal 2.
Definition at line 94 of file ClpLsqr.hpp.
The documentation for this class was generated from the following file:
- /tmp/CHiPPS-Alps-1.5.6/Clp/src/ClpLsqr.hpp
