Functions for turning an amatrix into an dh2matrix.
More...
|
pdclusterbasis | buildrow_amatrix_dclusterbasis (pcamatrix G, pcdblock b, pctruncmode tm, real eps) |
| Construct a directional row basis for an array matrix. More...
|
|
pdclusterbasis | buildcol_amatrix_dclusterbasis (pcamatrix G, pcdblock b, pctruncmode tm, real eps) |
| Construct a directional column basis for an array matrix. More...
|
|
void | collectdense_dh2matrix (pcamatrix a, pcdclusterbasis rb, uint rd, pcdclusterbasis cb, uint cd, pamatrix s) |
| Compute the optimal coupling matrix for a given matrix and given orthogonal row and column cluster bases. More...
|
|
void | projectdense_dh2matrix (pcamatrix a, pdh2matrix h2) |
| Fill a given dh2matrix object with the optimal coefficients for the approximation of a given matrix. More...
|
|
Functions for turning an amatrix into an dh2matrix.
The functions in this module can be used to convert array matrices into directional -matrices.
Construct a directional column basis for an array matrix.
- Parameters
-
G | Original matrix . |
b | Directional block tree. |
tm | Truncation mode. |
eps | Truncation accuracy. |
- Returns
- New column cluster basis.
Construct a directional row basis for an array matrix.
- Parameters
-
G | Original matrix . |
b | Directional block tree. |
tm | Truncation mode. |
eps | Truncation accuracy. |
- Returns
- New row cluster basis.
Compute the optimal coupling matrix for a given matrix and given orthogonal row and column cluster bases.
- Parameters
-
a | Source matrix, not in cluster ordering. |
rb | Orthogonal row cluster basis. |
rd | Direction for row basis. |
cb | Orthogonal column cluster basis. |
cd | Direction for column basis. |
s | Will be filled with optimal coefficients to approximate the given matrix. |
Fill a given dh2matrix object with the optimal coefficients for the approximation of a given matrix.
- Parameters
-
a | Source matrix, not in cluster ordering. |
h2 | Target matrix with orthogonal row and column cluster basis, nearfield and coupling matrices will be filled with coefficients approximation a . |