matrix_algorithms

wield.control.algorithms.statespace.dense.matrix_algorithms

Functions

GQR(matX, matY[, mshadowX, mshadowY, ...])	Implementation of COMPUTATION OF IRREDUCIBLE GENERALIZED STATE-SPACE REALIZATIONS ANDRAS VARGA using givens rotations.
QR(mat[, mshadow, Qapply, pivoting, method, ...])
norm1DcSq(u)
norm1DrSq(u)
swap_col(m, Cidx1, Cidx2)
swap_row(m, Cidx1, Cidx2)

Details

GQR(matX, matY, mshadowX=None, mshadowY=None, QZapply={}, overwrite=False, Rexact=False, zero_test=<function <lambda>>, tol=1e-09, shiftXcol=0, Ncols_end=None, Ncols_start=0, NHessenberg=None)

Implementation of COMPUTATION OF IRREDUCIBLE GENERALIZED STATE-SPACE REALIZATIONS ANDRAS VARGA using givens rotations.

It requires matY (E) to be in upper triangular form

it is very slow, but (mostly) numerically stable

TODO, add pivoting, TODO, make it use the U-T property on E better for speed TODO, make it output Q and Z to apply to aux matrices, perhaps use them on C

QR(mat, mshadow=None, Qapply={}, pivoting=False, method='Householder', overwrite=False, Rexact=False, zero_test=<function <lambda>>, select_pivot=None)

norm1DcSq(u)

norm1DrSq(u)

swap_col(m, Cidx1, Cidx2)

swap_row(m, Cidx1, Cidx2)