Hi Matthew,
The issue#360 in arnoldi
function is because of an extra multiplication of A
in the deflation.
The current (incorrect) arnoldi
calculates instead the first eigenvalues of A
and A-A^2
. That's why in issue#360 the two eigenvectors are identical and the eigenvalues
lambda0 = (0.598975,0)
lambda1 = (0.240204,0)
satisfy lambda1 = lambda0 - lambda0^2
.
A minor pull request has been submitted here.
However, even after this fix the arnoldi
function can only deal with Hermitian operator. Because the current deflation process cannot project to the correct subspace if the previously calculated eigenvectors are not orthogonal.
Ce