| 1 |
/* |
| 2 |
* Copyright (C) 2000-2004 Object Oriented Parallel Simulation Engine (OOPSE) project |
| 3 |
* |
| 4 |
* Contact: oopse@oopse.org |
| 5 |
* |
| 6 |
* This program is free software; you can redistribute it and/or |
| 7 |
* modify it under the terms of the GNU Lesser General Public License |
| 8 |
* as published by the Free Software Foundation; either version 2.1 |
| 9 |
* of the License, or (at your option) any later version. |
| 10 |
* All we ask is that proper credit is given for our work, which includes |
| 11 |
* - but is not limited to - adding the above copyright notice to the beginning |
| 12 |
* of your source code files, and to any copyright notice that you may distribute |
| 13 |
* with programs based on this work. |
| 14 |
* |
| 15 |
* This program is distributed in the hope that it will be useful, |
| 16 |
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 17 |
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 18 |
* GNU Lesser General Public License for more details. |
| 19 |
* |
| 20 |
* You should have received a copy of the GNU Lesser General Public License |
| 21 |
* along with this program; if not, write to the Free Software |
| 22 |
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. |
| 23 |
* |
| 24 |
*/ |
| 25 |
|
| 26 |
/** |
| 27 |
* @file SquareMatrix.hpp |
| 28 |
* @author Teng Lin |
| 29 |
* @date 10/11/2004 |
| 30 |
* @version 1.0 |
| 31 |
*/ |
| 32 |
#ifndef MATH_SQUAREMATRIX_HPP |
| 33 |
#define MATH_SQUAREMATRIX_HPP |
| 34 |
|
| 35 |
#include "math/RectMatrix.hpp" |
| 36 |
|
| 37 |
namespace oopse { |
| 38 |
|
| 39 |
/** |
| 40 |
* @class SquareMatrix SquareMatrix.hpp "math/SquareMatrix.hpp" |
| 41 |
* @brief A square matrix class |
| 42 |
* @template Real the element type |
| 43 |
* @template Dim the dimension of the square matrix |
| 44 |
*/ |
| 45 |
template<typename Real, int Dim> |
| 46 |
class SquareMatrix : public RectMatrix<Real, Dim, Dim> { |
| 47 |
public: |
| 48 |
|
| 49 |
/** default constructor */ |
| 50 |
SquareMatrix() { |
| 51 |
for (unsigned int i = 0; i < Dim; i++) |
| 52 |
for (unsigned int j = 0; j < Dim; j++) |
| 53 |
data_[i][j] = 0.0; |
| 54 |
} |
| 55 |
|
| 56 |
/** copy constructor */ |
| 57 |
SquareMatrix(const RectMatrix<Real, Dim, Dim>& m) : RectMatrix<Real, Dim, Dim>(m) { |
| 58 |
} |
| 59 |
|
| 60 |
/** copy assignment operator */ |
| 61 |
SquareMatrix<Real, Dim>& operator =(const RectMatrix<Real, Dim, Dim>& m) { |
| 62 |
RectMatrix<Real, Dim, Dim>::operator=(m); |
| 63 |
return *this; |
| 64 |
} |
| 65 |
|
| 66 |
/** Retunrs an identity matrix*/ |
| 67 |
|
| 68 |
static SquareMatrix<Real, Dim> identity() { |
| 69 |
SquareMatrix<Real, Dim> m; |
| 70 |
|
| 71 |
for (unsigned int i = 0; i < Dim; i++) |
| 72 |
for (unsigned int j = 0; j < Dim; j++) |
| 73 |
if (i == j) |
| 74 |
m(i, j) = 1.0; |
| 75 |
else |
| 76 |
m(i, j) = 0.0; |
| 77 |
|
| 78 |
return m; |
| 79 |
} |
| 80 |
|
| 81 |
/** Retunrs the inversion of this matrix. */ |
| 82 |
SquareMatrix<Real, Dim> inverse() { |
| 83 |
SquareMatrix<Real, Dim> result; |
| 84 |
|
| 85 |
return result; |
| 86 |
} |
| 87 |
|
| 88 |
/** Returns the determinant of this matrix. */ |
| 89 |
double determinant() const { |
| 90 |
double det; |
| 91 |
return det; |
| 92 |
} |
| 93 |
|
| 94 |
/** Returns the trace of this matrix. */ |
| 95 |
double trace() const { |
| 96 |
double tmp = 0; |
| 97 |
|
| 98 |
for (unsigned int i = 0; i < Dim ; i++) |
| 99 |
tmp += data_[i][i]; |
| 100 |
|
| 101 |
return tmp; |
| 102 |
} |
| 103 |
|
| 104 |
/** Tests if this matrix is symmetrix. */ |
| 105 |
bool isSymmetric() const { |
| 106 |
for (unsigned int i = 0; i < Dim - 1; i++) |
| 107 |
for (unsigned int j = i; j < Dim; j++) |
| 108 |
if (fabs(data_[i][j] - data_[j][i]) > oopse::epsilon) |
| 109 |
return false; |
| 110 |
|
| 111 |
return true; |
| 112 |
} |
| 113 |
|
| 114 |
/** Tests if this matrix is orthogonal. */ |
| 115 |
bool isOrthogonal() { |
| 116 |
SquareMatrix<Real, Dim> tmp; |
| 117 |
|
| 118 |
tmp = *this * transpose(); |
| 119 |
|
| 120 |
return tmp.isDiagonal(); |
| 121 |
} |
| 122 |
|
| 123 |
/** Tests if this matrix is diagonal. */ |
| 124 |
bool isDiagonal() const { |
| 125 |
for (unsigned int i = 0; i < Dim ; i++) |
| 126 |
for (unsigned int j = 0; j < Dim; j++) |
| 127 |
if (i !=j && fabs(data_[i][j]) > oopse::epsilon) |
| 128 |
return false; |
| 129 |
|
| 130 |
return true; |
| 131 |
} |
| 132 |
|
| 133 |
/** Tests if this matrix is the unit matrix. */ |
| 134 |
bool isUnitMatrix() const { |
| 135 |
if (!isDiagonal()) |
| 136 |
return false; |
| 137 |
|
| 138 |
for (unsigned int i = 0; i < Dim ; i++) |
| 139 |
if (fabs(data_[i][i] - 1) > oopse::epsilon) |
| 140 |
return false; |
| 141 |
|
| 142 |
return true; |
| 143 |
} |
| 144 |
|
| 145 |
void diagonalize() { |
| 146 |
jacobi(m, eigenValues, ortMat); |
| 147 |
} |
| 148 |
|
| 149 |
/** |
| 150 |
* Finds the eigenvalues and eigenvectors of a symmetric matrix |
| 151 |
* @param eigenvals a reference to a vector3 where the |
| 152 |
* eigenvalues will be stored. The eigenvalues are ordered so |
| 153 |
* that eigenvals[0] <= eigenvals[1] <= eigenvals[2]. |
| 154 |
* @return an orthogonal matrix whose ith column is an |
| 155 |
* eigenvector for the eigenvalue eigenvals[i] |
| 156 |
*/ |
| 157 |
SquareMatrix<Real, Dim> findEigenvectors(Vector<Real, Dim>& eigenValues) { |
| 158 |
SquareMatrix<Real, Dim> ortMat; |
| 159 |
|
| 160 |
if ( !isSymmetric()){ |
| 161 |
throw(); |
| 162 |
} |
| 163 |
|
| 164 |
SquareMatrix<Real, Dim> m(*this); |
| 165 |
jacobi(m, eigenValues, ortMat); |
| 166 |
|
| 167 |
return ortMat; |
| 168 |
} |
| 169 |
/** |
| 170 |
* Jacobi iteration routines for computing eigenvalues/eigenvectors of |
| 171 |
* real symmetric matrix |
| 172 |
* |
| 173 |
* @return true if success, otherwise return false |
| 174 |
* @param a source matrix |
| 175 |
* @param w output eigenvalues |
| 176 |
* @param v output eigenvectors |
| 177 |
*/ |
| 178 |
bool jacobi(const SquareMatrix<Real, Dim>& a, Vector<Real, Dim>& w, |
| 179 |
SquareMatrix<Real, Dim>& v); |
| 180 |
};//end SquareMatrix |
| 181 |
|
| 182 |
|
| 183 |
#define ROT(a,i,j,k,l) g=a(i, j);h=a(k, l);a(i, j)=g-s*(h+g*tau);a(k, l)=h+s*(g-h*tau) |
| 184 |
#define MAX_ROTATIONS 60 |
| 185 |
|
| 186 |
template<typename Real, int Dim> |
| 187 |
bool SquareMatrix<Real, Dim>::jacobi(const SquareMatrix<Real, Dim>& a, Vector<Real, Dim>& w, |
| 188 |
SquareMatrix<Real, Dim>& v) { |
| 189 |
const int N = Dim; |
| 190 |
int i, j, k, iq, ip; |
| 191 |
double tresh, theta, tau, t, sm, s, h, g, c; |
| 192 |
double tmp; |
| 193 |
Vector<Real, Dim> b, z; |
| 194 |
|
| 195 |
// initialize |
| 196 |
for (ip=0; ip<N; ip++) { |
| 197 |
for (iq=0; iq<N; iq++) |
| 198 |
v(ip, iq) = 0.0; |
| 199 |
v(ip, ip) = 1.0; |
| 200 |
} |
| 201 |
|
| 202 |
for (ip=0; ip<N; ip++) { |
| 203 |
b(ip) = w(ip) = a(ip, ip); |
| 204 |
z(ip) = 0.0; |
| 205 |
} |
| 206 |
|
| 207 |
// begin rotation sequence |
| 208 |
for (i=0; i<MAX_ROTATIONS; i++) { |
| 209 |
sm = 0.0; |
| 210 |
for (ip=0; ip<2; ip++) { |
| 211 |
for (iq=ip+1; iq<N; iq++) |
| 212 |
sm += fabs(a(ip, iq)); |
| 213 |
} |
| 214 |
|
| 215 |
if (sm == 0.0) |
| 216 |
break; |
| 217 |
|
| 218 |
if (i < 4) |
| 219 |
tresh = 0.2*sm/(9); |
| 220 |
else |
| 221 |
tresh = 0.0; |
| 222 |
|
| 223 |
for (ip=0; ip<2; ip++) { |
| 224 |
for (iq=ip+1; iq<N; iq++) { |
| 225 |
g = 100.0*fabs(a(ip, iq)); |
| 226 |
if (i > 4 && (fabs(w(ip))+g) == fabs(w(ip)) |
| 227 |
&& (fabs(w(iq))+g) == fabs(w(iq))) { |
| 228 |
a(ip, iq) = 0.0; |
| 229 |
} else if (fabs(a(ip, iq)) > tresh) { |
| 230 |
h = w(iq) - w(ip); |
| 231 |
if ( (fabs(h)+g) == fabs(h)) { |
| 232 |
t = (a(ip, iq)) / h; |
| 233 |
} else { |
| 234 |
theta = 0.5*h / (a(ip, iq)); |
| 235 |
t = 1.0 / (fabs(theta)+sqrt(1.0+theta*theta)); |
| 236 |
|
| 237 |
if (theta < 0.0) |
| 238 |
t = -t; |
| 239 |
} |
| 240 |
|
| 241 |
c = 1.0 / sqrt(1+t*t); |
| 242 |
s = t*c; |
| 243 |
tau = s/(1.0+c); |
| 244 |
h = t*a(ip, iq); |
| 245 |
z(ip) -= h; |
| 246 |
z(iq) += h; |
| 247 |
w(ip) -= h; |
| 248 |
w(iq) += h; |
| 249 |
a(ip, iq)=0.0; |
| 250 |
|
| 251 |
for (j=0;j<ip-1;j++) |
| 252 |
ROT(a,j,ip,j,iq); |
| 253 |
|
| 254 |
for (j=ip+1;j<iq-1;j++) |
| 255 |
ROT(a,ip,j,j,iq); |
| 256 |
|
| 257 |
for (j=iq+1; j<N; j++) |
| 258 |
ROT(a,ip,j,iq,j); |
| 259 |
for (j=0; j<N; j++) |
| 260 |
ROT(v,j,ip,j,iq); |
| 261 |
} |
| 262 |
} |
| 263 |
}//for (ip=0; ip<2; ip++) |
| 264 |
|
| 265 |
for (ip=0; ip<N; ip++) { |
| 266 |
b(ip) += z(ip); |
| 267 |
w(ip) = b(ip); |
| 268 |
z(ip) = 0.0; |
| 269 |
} |
| 270 |
|
| 271 |
} // end for (i=0; i<MAX_ROTATIONS; i++) |
| 272 |
|
| 273 |
if ( i >= MAX_ROTATIONS ) |
| 274 |
return false; |
| 275 |
|
| 276 |
// sort eigenfunctions |
| 277 |
for (j=0; j<N; j++) { |
| 278 |
k = j; |
| 279 |
tmp = w(k); |
| 280 |
for (i=j; i<N; i++) { |
| 281 |
if (w(i) >= tmp) { |
| 282 |
k = i; |
| 283 |
tmp = w(k); |
| 284 |
} |
| 285 |
} |
| 286 |
|
| 287 |
if (k != j) { |
| 288 |
w(k) = w(j); |
| 289 |
w(j) = tmp; |
| 290 |
for (i=0; i<N; i++) { |
| 291 |
tmp = v(i, j); |
| 292 |
v(i, j) = v(i, k); |
| 293 |
v(i, k) = tmp; |
| 294 |
} |
| 295 |
} |
| 296 |
} |
| 297 |
|
| 298 |
// insure eigenvector consistency (i.e., Jacobi can compute |
| 299 |
// vectors that are negative of one another (.707,.707,0) and |
| 300 |
// (-.707,-.707,0). This can reek havoc in |
| 301 |
// hyperstreamline/other stuff. We will select the most |
| 302 |
// positive eigenvector. |
| 303 |
int numPos; |
| 304 |
for (j=0; j<N; j++) { |
| 305 |
for (numPos=0, i=0; i<N; i++) if ( v(i, j) >= 0.0 ) numPos++; |
| 306 |
if ( numPos < 2 ) for(i=0; i<N; i++) v(i, j) *= -1.0; |
| 307 |
} |
| 308 |
|
| 309 |
return true; |
| 310 |
} |
| 311 |
|
| 312 |
#undef ROT |
| 313 |
#undef MAX_ROTATIONS |
| 314 |
|
| 315 |
} |
| 316 |
|
| 317 |
#endif //MATH_SQUAREMATRIX_HPP |
