Files
openfoam/src/OpenFOAM/matrices/scalarMatrices/SVD/SVD.C
henry 318b71e206 Made Matrix generic base-class for SquareMatrix and RectangularMatrix by introducing
the type it will become as a template argument.

Brought everything into line with this change.
2008-09-28 22:50:57 +01:00

404 lines
10 KiB
C

/*---------------------------------------------------------------------------*\
========= |
\\ / F ield | OpenFOAM: The Open Source CFD Toolbox
\\ / O peration |
\\ / A nd | Copyright (C) 1991-2005 OpenCFD Ltd.
\\/ M anipulation |
-------------------------------------------------------------------------------
License
This file is part of OpenFOAM.
OpenFOAM is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2 of the License, or (at your
option) any later version.
OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License
along with OpenFOAM; if not, write to the Free Software Foundation,
Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
\*---------------------------------------------------------------------------*/
#include "SVD.H"
#include "scalarList.H"
#include "scalarMatrices.H"
#include "ListOps.H"
// * * * * * * * * * * * * * * * * Constructor * * * * * * * * * * * * * * //
Foam::SVD::SVD(const scalarRectangularMatrix& A, const scalar minCondition)
:
U_(A),
V_(A.m(), A.m()),
S_(A.m()),
VSinvUt_(A.m(), A.n()),
nZeros_(0)
{
// SVDcomp to find U_, V_ and S_ - the singular values
const label Um = U_.m();
const label Un = U_.n();
scalarList rv1(Um);
scalar g = 0;
scalar scale = 0;
scalar s = 0;
scalar anorm = 0;
label l = 0;
for (label i = 0; i < Um; i++)
{
l = i+2;
rv1[i] = scale*g;
g = s = scale = 0;
if (i < Un)
{
for (label k = i; k < Un; k++)
{
scale += mag(U_[k][i]);
}
if (scale != 0)
{
for (label k = i; k < Un; k++)
{
U_[k][i] /= scale;
s += U_[k][i]*U_[k][i];
}
scalar f = U_[i][i];
g = -sign(Foam::sqrt(s), f);
scalar h = f*g - s;
U_[i][i] = f - g;
for (label j = l-1; j < Um; j++)
{
s = 0;
for (label k = i; k < Un; k++)
{
s += U_[k][i]*U_[k][j];
}
f = s/h;
for (label k = i; k < A.n(); k++)
{
U_[k][j] += f*U_[k][i];
}
}
for (label k = i; k < Un; k++)
{
U_[k][i] *= scale;
}
}
}
S_[i] = scale*g;
g = s = scale = 0;
if (i+1 <= Un && i != Um)
{
for (label k = l-1; k < Um; k++)
{
scale += mag(U_[i][k]);
}
if (scale != 0)
{
for (label k=l-1; k < Um; k++)
{
U_[i][k] /= scale;
s += U_[i][k]*U_[i][k];
}
scalar f = U_[i][l-1];
g = -sign(Foam::sqrt(s),f);
scalar h = f*g - s;
U_[i][l-1] = f - g;
for (label k = l-1; k < Um; k++)
{
rv1[k] = U_[i][k]/h;
}
for (label j = l-1; j < Un; j++)
{
s = 0;
for (label k = l-1; k < Um; k++)
{
s += U_[j][k]*U_[i][k];
}
for (label k = l-1; k < Um; k++)
{
U_[j][k] += s*rv1[k];
}
}
for (label k = l-1; k < Um; k++)
{
U_[i][k] *= scale;
}
}
}
anorm = max(anorm, mag(S_[i]) + mag(rv1[i]));
}
for (label i = Um-1; i >= 0; i--)
{
if (i < Um-1)
{
if (g != 0)
{
for (label j = l; j < Um; j++)
{
V_[j][i] = (U_[i][j]/U_[i][l])/g;
}
for (label j=l; j < Um; j++)
{
s = 0;
for (label k = l; k < Um; k++)
{
s += U_[i][k]*V_[k][j];
}
for (label k = l; k < Um; k++)
{
V_[k][j] += s*V_[k][i];
}
}
}
for (label j = l; j < Um;j++)
{
V_[i][j] = V_[j][i] = 0.0;
}
}
V_[i][i] = 1;
g = rv1[i];
l = i;
}
for (label i = min(Um, Un) - 1; i >= 0; i--)
{
l = i+1;
g = S_[i];
for (label j = l; j < Um; j++)
{
U_[i][j] = 0.0;
}
if (g != 0)
{
g = 1.0/g;
for (label j = l; j < Um; j++)
{
s = 0;
for (label k = l; k < Un; k++)
{
s += U_[k][i]*U_[k][j];
}
scalar f = (s/U_[i][i])*g;
for (label k = i; k < Un; k++)
{
U_[k][j] += f*U_[k][i];
}
}
for (label j = i; j < Un; j++)
{
U_[j][i] *= g;
}
}
else
{
for (label j = i; j < Un; j++)
{
U_[j][i] = 0.0;
}
}
++U_[i][i];
}
for (label k = Um-1; k >= 0; k--)
{
for (label its = 0; its < 35; its++)
{
bool flag = true;
label nm;
for (l = k; l >= 0; l--)
{
nm = l-1;
if (mag(rv1[l]) + anorm == anorm)
{
flag = false;
break;
}
if (mag(S_[nm]) + anorm == anorm) break;
}
if (flag)
{
scalar c = 0.0;
s = 1.0;
for (label i = l-1; i < k+1; i++)
{
scalar f = s*rv1[i];
rv1[i] = c*rv1[i];
if (mag(f) + anorm == anorm) break;
g = S_[i];
scalar h = sqrtSumSqr(f, g);
S_[i] = h;
h = 1.0/h;
c = g*h;
s = -f*h;
for (label j = 0; j < Un; j++)
{
scalar y = U_[j][nm];
scalar z = U_[j][i];
U_[j][nm] = y*c + z*s;
U_[j][i] = z*c - y*s;
}
}
}
scalar z = S_[k];
if (l == k)
{
if (z < 0.0)
{
S_[k] = -z;
for (label j = 0; j < Um; j++) V_[j][k] = -V_[j][k];
}
break;
}
if (its == 34)
{
WarningIn
(
"SVD::SVD"
"(scalarRectangularMatrix& A, const scalar minCondition)"
) << "no convergence in 35 SVD iterations"
<< endl;
}
scalar x = S_[l];
nm = k-1;
scalar y = S_[nm];
g = rv1[nm];
scalar h = rv1[k];
scalar f = ((y - z)*(y + z) + (g - h)*(g + h))/(2.0*h*y);
g = sqrtSumSqr(f, 1.0);
f = ((x - z)*(x + z) + h*((y/(f + sign(g, f))) - h))/x;
scalar c = 1.0;
s = 1.0;
for (label j = l; j <= nm; j++)
{
label i = j + 1;
g = rv1[i];
y = S_[i];
h = s*g;
g = c*g;
scalar z = sqrtSumSqr(f, h);
rv1[j] = z;
c = f/z;
s = h/z;
f = x*c + g*s;
g = g*c - x*s;
h = y*s;
y *= c;
for (label jj = 0; jj < Um; jj++)
{
x = V_[jj][j];
z = V_[jj][i];
V_[jj][j] = x*c + z*s;
V_[jj][i] = z*c - x*s;
}
z = sqrtSumSqr(f, h);
S_[j] = z;
if (z)
{
z = 1.0/z;
c = f*z;
s = h*z;
}
f = c*g + s*y;
x = c*y - s*g;
for (label jj=0; jj < Un; jj++)
{
y = U_[jj][j];
z = U_[jj][i];
U_[jj][j] = y*c + z*s;
U_[jj][i] = z*c - y*s;
}
}
rv1[l] = 0.0;
rv1[k] = f;
S_[k] = x;
}
}
// zero singular values that are less than minCondition*maxS
const scalar minS = minCondition*S_[findMax(S_)];
for (label i = 0; i < S_.size(); i++)
{
if (S_[i] <= minS)
{
//Info << "Removing " << S_[i] << " < " << minS << endl;
S_[i] = 0;
nZeros_++;
}
}
// now multiply out to find the pseudo inverse of A, VSinvUt_
multiply(VSinvUt_, V_, inv(S_), U_.T());
// test SVD
/*scalarRectangularMatrix SVDA(A.n(), A.m());
multiply(SVDA, U_, S_, transpose(V_));
scalar maxDiff = 0;
scalar diff = 0;
for(label i = 0; i < A.n(); i++)
{
for(label j = 0; j < A.m(); j++)
{
diff = mag(A[i][j] - SVDA[i][j]);
if (diff > maxDiff) maxDiff = diff;
}
}
Info << "Maximum discrepancy between A and svd(A) = " << maxDiff << endl;
if (maxDiff > 4)
{
Info << "singular values " << S_ << endl;
}
*/
}
// ************************************************************************* //