IncompleteLUT.h
Go to the documentation of this file.
1 // This file is part of Eigen, a lightweight C++ template library
2 // for linear algebra.
3 //
4 // Copyright (C) 2012 Désiré Nuentsa-Wakam <desire.nuentsa_wakam@inria.fr>
5 //
6 // This Source Code Form is subject to the terms of the Mozilla
7 // Public License v. 2.0. If a copy of the MPL was not distributed
8 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
9 
10 #ifndef EIGEN_INCOMPLETE_LUT_H
11 #define EIGEN_INCOMPLETE_LUT_H
12 
13 
14 namespace Eigen {
15 
16 namespace internal {
17 
27 template <typename VectorV, typename VectorI, typename Index>
28 Index QuickSplit(VectorV &row, VectorI &ind, Index ncut)
29 {
30  typedef typename VectorV::RealScalar RealScalar;
31  using std::swap;
32  using std::abs;
33  Index mid;
34  Index n = row.size(); /* length of the vector */
35  Index first, last ;
36 
37  ncut--; /* to fit the zero-based indices */
38  first = 0;
39  last = n-1;
40  if (ncut < first || ncut > last ) return 0;
41 
42  do {
43  mid = first;
44  RealScalar abskey = abs(row(mid));
45  for (Index j = first + 1; j <= last; j++) {
46  if ( abs(row(j)) > abskey) {
47  ++mid;
48  swap(row(mid), row(j));
49  swap(ind(mid), ind(j));
50  }
51  }
52  /* Interchange for the pivot element */
53  swap(row(mid), row(first));
54  swap(ind(mid), ind(first));
55 
56  if (mid > ncut) last = mid - 1;
57  else if (mid < ncut ) first = mid + 1;
58  } while (mid != ncut );
59 
60  return 0; /* mid is equal to ncut */
61 }
62 
63 }// end namespace internal
64 
95 template <typename _Scalar>
97 {
98  typedef _Scalar Scalar;
103  typedef typename FactorType::Index Index;
104 
105  public:
107 
109  : m_droptol(NumTraits<Scalar>::dummy_precision()), m_fillfactor(10),
110  m_analysisIsOk(false), m_factorizationIsOk(false), m_isInitialized(false)
111  {}
112 
113  template<typename MatrixType>
114  IncompleteLUT(const MatrixType& mat, const RealScalar& droptol=NumTraits<Scalar>::dummy_precision(), int fillfactor = 10)
115  : m_droptol(droptol),m_fillfactor(fillfactor),
116  m_analysisIsOk(false),m_factorizationIsOk(false),m_isInitialized(false)
117  {
118  eigen_assert(fillfactor != 0);
119  compute(mat);
120  }
121 
122  Index rows() const { return m_lu.rows(); }
123 
124  Index cols() const { return m_lu.cols(); }
125 
132  {
133  eigen_assert(m_isInitialized && "IncompleteLUT is not initialized.");
134  return m_info;
135  }
136 
137  template<typename MatrixType>
138  void analyzePattern(const MatrixType& amat);
139 
140  template<typename MatrixType>
141  void factorize(const MatrixType& amat);
142 
148  template<typename MatrixType>
149  IncompleteLUT<Scalar>& compute(const MatrixType& amat)
150  {
151  analyzePattern(amat);
152  factorize(amat);
153  m_isInitialized = m_factorizationIsOk;
154  return *this;
155  }
156 
157  void setDroptol(const RealScalar& droptol);
158  void setFillfactor(int fillfactor);
159 
160  template<typename Rhs, typename Dest>
161  void _solve(const Rhs& b, Dest& x) const
162  {
163  x = m_Pinv * b;
164  x = m_lu.template triangularView<UnitLower>().solve(x);
165  x = m_lu.template triangularView<Upper>().solve(x);
166  x = m_P * x;
167  }
168 
169  template<typename Rhs> inline const internal::solve_retval<IncompleteLUT, Rhs>
170  solve(const MatrixBase<Rhs>& b) const
171  {
172  eigen_assert(m_isInitialized && "IncompleteLUT is not initialized.");
173  eigen_assert(cols()==b.rows()
174  && "IncompleteLUT::solve(): invalid number of rows of the right hand side matrix b");
175  return internal::solve_retval<IncompleteLUT, Rhs>(*this, b.derived());
176  }
177 
178 protected:
179 
181  struct keep_diag {
182  inline bool operator() (const Index& row, const Index& col, const Scalar&) const
183  {
184  return row!=col;
185  }
186  };
187 
188 protected:
189 
190  FactorType m_lu;
191  RealScalar m_droptol;
197  PermutationMatrix<Dynamic,Dynamic,Index> m_P; // Fill-reducing permutation
199 };
200 
205 template<typename Scalar>
206 void IncompleteLUT<Scalar>::setDroptol(const RealScalar& droptol)
207 {
208  this->m_droptol = droptol;
209 }
210 
215 template<typename Scalar>
217 {
218  this->m_fillfactor = fillfactor;
219 }
220 
221 template <typename Scalar>
222 template<typename _MatrixType>
223 void IncompleteLUT<Scalar>::analyzePattern(const _MatrixType& amat)
224 {
225  // Compute the Fill-reducing permutation
228  // Symmetrize the pattern
229  // FIXME for a matrix with nearly symmetric pattern, mat2+mat1 is the appropriate choice.
230  // on the other hand for a really non-symmetric pattern, mat2*mat1 should be prefered...
231  SparseMatrix<Scalar,ColMajor, Index> AtA = mat2 + mat1;
232  AtA.prune(keep_diag());
233  internal::minimum_degree_ordering<Scalar, Index>(AtA, m_P); // Then compute the AMD ordering...
234 
235  m_Pinv = m_P.inverse(); // ... and the inverse permutation
236 
237  m_analysisIsOk = true;
238 }
239 
240 template <typename Scalar>
241 template<typename _MatrixType>
242 void IncompleteLUT<Scalar>::factorize(const _MatrixType& amat)
243 {
244  using std::sqrt;
245  using std::swap;
246  using std::abs;
247 
248  eigen_assert((amat.rows() == amat.cols()) && "The factorization should be done on a square matrix");
249  Index n = amat.cols(); // Size of the matrix
250  m_lu.resize(n,n);
251  // Declare Working vectors and variables
252  Vector u(n) ; // real values of the row -- maximum size is n --
253  VectorXi ju(n); // column position of the values in u -- maximum size is n
254  VectorXi jr(n); // Indicate the position of the nonzero elements in the vector u -- A zero location is indicated by -1
255 
256  // Apply the fill-reducing permutation
257  eigen_assert(m_analysisIsOk && "You must first call analyzePattern()");
259  mat = amat.twistedBy(m_Pinv);
260 
261  // Initialization
262  jr.fill(-1);
263  ju.fill(0);
264  u.fill(0);
265 
266  // number of largest elements to keep in each row:
267  Index fill_in = static_cast<Index> (amat.nonZeros()*m_fillfactor)/n+1;
268  if (fill_in > n) fill_in = n;
269 
270  // number of largest nonzero elements to keep in the L and the U part of the current row:
271  Index nnzL = fill_in/2;
272  Index nnzU = nnzL;
273  m_lu.reserve(n * (nnzL + nnzU + 1));
274 
275  // global loop over the rows of the sparse matrix
276  for (Index ii = 0; ii < n; ii++)
277  {
278  // 1 - copy the lower and the upper part of the row i of mat in the working vector u
279 
280  Index sizeu = 1; // number of nonzero elements in the upper part of the current row
281  Index sizel = 0; // number of nonzero elements in the lower part of the current row
282  ju(ii) = ii;
283  u(ii) = 0;
284  jr(ii) = ii;
285  RealScalar rownorm = 0;
286 
287  typename FactorType::InnerIterator j_it(mat, ii); // Iterate through the current row ii
288  for (; j_it; ++j_it)
289  {
290  Index k = j_it.index();
291  if (k < ii)
292  {
293  // copy the lower part
294  ju(sizel) = k;
295  u(sizel) = j_it.value();
296  jr(k) = sizel;
297  ++sizel;
298  }
299  else if (k == ii)
300  {
301  u(ii) = j_it.value();
302  }
303  else
304  {
305  // copy the upper part
306  Index jpos = ii + sizeu;
307  ju(jpos) = k;
308  u(jpos) = j_it.value();
309  jr(k) = jpos;
310  ++sizeu;
311  }
312  rownorm += numext::abs2(j_it.value());
313  }
314 
315  // 2 - detect possible zero row
316  if(rownorm==0)
317  {
318  m_info = NumericalIssue;
319  return;
320  }
321  // Take the 2-norm of the current row as a relative tolerance
322  rownorm = sqrt(rownorm);
323 
324  // 3 - eliminate the previous nonzero rows
325  Index jj = 0;
326  Index len = 0;
327  while (jj < sizel)
328  {
329  // In order to eliminate in the correct order,
330  // we must select first the smallest column index among ju(jj:sizel)
331  Index k;
332  Index minrow = ju.segment(jj,sizel-jj).minCoeff(&k); // k is relative to the segment
333  k += jj;
334  if (minrow != ju(jj))
335  {
336  // swap the two locations
337  Index j = ju(jj);
338  swap(ju(jj), ju(k));
339  jr(minrow) = jj; jr(j) = k;
340  swap(u(jj), u(k));
341  }
342  // Reset this location
343  jr(minrow) = -1;
344 
345  // Start elimination
346  typename FactorType::InnerIterator ki_it(m_lu, minrow);
347  while (ki_it && ki_it.index() < minrow) ++ki_it;
348  eigen_internal_assert(ki_it && ki_it.col()==minrow);
349  Scalar fact = u(jj) / ki_it.value();
350 
351  // drop too small elements
352  if(abs(fact) <= m_droptol)
353  {
354  jj++;
355  continue;
356  }
357 
358  // linear combination of the current row ii and the row minrow
359  ++ki_it;
360  for (; ki_it; ++ki_it)
361  {
362  Scalar prod = fact * ki_it.value();
363  Index j = ki_it.index();
364  Index jpos = jr(j);
365  if (jpos == -1) // fill-in element
366  {
367  Index newpos;
368  if (j >= ii) // dealing with the upper part
369  {
370  newpos = ii + sizeu;
371  sizeu++;
372  eigen_internal_assert(sizeu<=n);
373  }
374  else // dealing with the lower part
375  {
376  newpos = sizel;
377  sizel++;
378  eigen_internal_assert(sizel<=ii);
379  }
380  ju(newpos) = j;
381  u(newpos) = -prod;
382  jr(j) = newpos;
383  }
384  else
385  u(jpos) -= prod;
386  }
387  // store the pivot element
388  u(len) = fact;
389  ju(len) = minrow;
390  ++len;
391 
392  jj++;
393  } // end of the elimination on the row ii
394 
395  // reset the upper part of the pointer jr to zero
396  for(Index k = 0; k <sizeu; k++) jr(ju(ii+k)) = -1;
397 
398  // 4 - partially sort and insert the elements in the m_lu matrix
399 
400  // sort the L-part of the row
401  sizel = len;
402  len = (std::min)(sizel, nnzL);
403  typename Vector::SegmentReturnType ul(u.segment(0, sizel));
404  typename VectorXi::SegmentReturnType jul(ju.segment(0, sizel));
405  internal::QuickSplit(ul, jul, len);
406 
407  // store the largest m_fill elements of the L part
408  m_lu.startVec(ii);
409  for(Index k = 0; k < len; k++)
410  m_lu.insertBackByOuterInnerUnordered(ii,ju(k)) = u(k);
411 
412  // store the diagonal element
413  // apply a shifting rule to avoid zero pivots (we are doing an incomplete factorization)
414  if (u(ii) == Scalar(0))
415  u(ii) = sqrt(m_droptol) * rownorm;
416  m_lu.insertBackByOuterInnerUnordered(ii, ii) = u(ii);
417 
418  // sort the U-part of the row
419  // apply the dropping rule first
420  len = 0;
421  for(Index k = 1; k < sizeu; k++)
422  {
423  if(abs(u(ii+k)) > m_droptol * rownorm )
424  {
425  ++len;
426  u(ii + len) = u(ii + k);
427  ju(ii + len) = ju(ii + k);
428  }
429  }
430  sizeu = len + 1; // +1 to take into account the diagonal element
431  len = (std::min)(sizeu, nnzU);
432  typename Vector::SegmentReturnType uu(u.segment(ii+1, sizeu-1));
433  typename VectorXi::SegmentReturnType juu(ju.segment(ii+1, sizeu-1));
434  internal::QuickSplit(uu, juu, len);
435 
436  // store the largest elements of the U part
437  for(Index k = ii + 1; k < ii + len; k++)
438  m_lu.insertBackByOuterInnerUnordered(ii,ju(k)) = u(k);
439  }
440 
441  m_lu.finalize();
442  m_lu.makeCompressed();
443 
444  m_factorizationIsOk = true;
445  m_info = Success;
446 }
447 
448 namespace internal {
449 
450 template<typename _MatrixType, typename Rhs>
451 struct solve_retval<IncompleteLUT<_MatrixType>, Rhs>
452  : solve_retval_base<IncompleteLUT<_MatrixType>, Rhs>
453 {
455  EIGEN_MAKE_SOLVE_HELPERS(Dec,Rhs)
456 
457  template<typename Dest> void evalTo(Dest& dst) const
458  {
459  dec()._solve(rhs(),dst);
460  }
461 };
462 
463 } // end namespace internal
464 
465 } // end namespace Eigen
466 
467 #endif // EIGEN_INCOMPLETE_LUT_H
FactorType::Index Index
VectorBlock< Derived > SegmentReturnType
Definition: BlockMethods.h:33
PermutationMatrix< Dynamic, Dynamic, Index > m_Pinv
Definition: LDLT.h:16
const internal::solve_retval< IncompleteLUT, Rhs > solve(const MatrixBase< Rhs > &b) const
Holds information about the various numeric (i.e. scalar) types allowed by Eigen. ...
Definition: NumTraits.h:88
#define eigen_internal_assert(x)
IncompleteLUT(const MatrixType &mat, const RealScalar &droptol=NumTraits< Scalar >::dummy_precision(), int fillfactor=10)
EIGEN_STRONG_INLINE const CwiseUnaryOp< internal::scalar_abs2_op< Scalar >, const Derived > abs2() const
SparseMatrix< Scalar, ColMajor > PermutType
Index cols() const
void setDroptol(const RealScalar &droptol)
EIGEN_STRONG_INLINE const CwiseUnaryOp< internal::scalar_abs_op< Scalar >, const Derived > abs() const
SparseSymmetricPermutationProduct< Derived, Upper|Lower > twistedBy(const PermutationMatrix< Dynamic, Dynamic, Index > &perm) const
PermutationMatrix< Dynamic, Dynamic, Index > m_P
void prune(const Scalar &reference, const RealScalar &epsilon=NumTraits< RealScalar >::dummy_precision())
Definition: SparseMatrix.h:490
void setFillfactor(int fillfactor)
IncompleteLUT< Scalar > & compute(const MatrixType &amat)
void analyzePattern(const MatrixType &amat)
Transpose< Derived > transpose()
ComputationInfo info() const
Reports whether previous computation was successful.
Incomplete LU factorization with dual-threshold strategy.
Definition: IncompleteLUT.h:96
internal::traits< Derived >::Index Index
Definition: EigenBase.h:31
Matrix< Scalar, Dynamic, 1 > Vector
Index QuickSplit(VectorV &row, VectorI &ind, Index ncut)
Definition: IncompleteLUT.h:28
SparseMatrix< Scalar, RowMajor > FactorType
Matrix< Scalar, Dynamic, Dynamic > MatrixType
ComputationInfo m_info
RowXpr row(Index i)
Definition: BlockMethods.h:725
void factorize(const MatrixType &amat)
void _solve(const Rhs &b, Dest &x) const
#define EIGEN_MAKE_SOLVE_HELPERS(DecompositionType, Rhs)
Definition: Solve.h:61
const CwiseUnaryOp< internal::scalar_sqrt_op< Scalar >, const Derived > sqrt() const
ColXpr col(Index i)
Definition: BlockMethods.h:708
#define eigen_assert(x)
ComputationInfo
Definition: Constants.h:374
Index rows() const
Base class for all dense matrices, vectors, and expressions.
Definition: MatrixBase.h:48
NumTraits< Scalar >::Real RealScalar
Definition: IncompleteLUT.h:99


tuw_aruco
Author(s): Lukas Pfeifhofer
autogenerated on Mon Jun 10 2019 15:40:51