You can download the new version of linearlagebra from
http://www.unk.edu/facstaff/profiles/willisb/
Look for linearalgebra(3).zip. I thank alert users Anita Kean,
Milan Lukic, and Dan Stanger for bug reports and suggestions on
earlier versions.
New functions: lu_factor, lu_get_factors, lu_backsub, kronnecker_product,
hilbert_matrix, vandermonde_matrix, mat_norm, mat_unblocker, conjugate
(an alternative to /eigen/conjugate) matrixp (extension of existing
matrixp),
listp (extension of existing listp), zeromatrixp, and several
miscellaneous functions.
Notes:
(1) For numeric matrices, the lu_factor code uses partial pivoting. The
code isn't fast, but it does work for symbolic, big float, and floating
point
(complex float too) matrices. Actually, you should be able to get the LU
code
to work with matrices with entries of any data type (even ones not known to
Maxima) by defining the arithmetic operations and a new mfield
structure (see lu.lisp).
(2) The eigen/conguate function only does the substitution %i --> -%i.
The simplifying conjugate function in linearalgebra is (I hope) somewhat
smarter.
(3) linearalgebra now contains my matrixexp code (functions for the
matrix spectral decomposition.)
As always, send me your suggestions.
Regards,
Barton