First of all, big thanks to all who replied
trying to help. I deeply respect your effort,
even if I'm unable to use the product effectively.
@Stavros: Actually, the matrix I need solved is
somewhat more complicated. What I posted (and
what still can't be solved, anyway) was the
result of (over)simplification, symbolic
variables assumed to be fractions of A, several
parameters assumed to be 'equal' etc.
@Richard: I called it 'simple' because it can
easily get quite un-simple. Posted (simplified)
matrix was a Hamiltonian with
cross-anharmonicities limited to neighbours
third-removed. Hence lots of sparseness, uniform
off-diagonal elements etc.
@Leo: The degree might get even higher. See
above. As for the charpoly output -- well, the
structure of eigenvalues output was ideal for my
purposes, so I'm shying from digging in
charpoly's. Of course, if I must, I must. I'd
still prefer eigenvalues output, though.
@All: Can I generate the series approximation
(in symbolic variables) in Maxima, for the
exponential/matrix exponential? Preferrably,
with adjustable cutoff (up to first, second
powers)? Sections 26 and 28 of manual seem to be
of no help.
-Yury