I have a problem:
for square symmetric matrix with eigenvalues of multiplicity > 1 Jacobi method(11.1) can't find right eigenvectors. At least they don't coincide with one Matlab gave me as a result.
Eigenvalues do coincide.
In other cases everything is ok.
I use the code, written in the book.
Is there any explanation?

Denis,

I tried computing the eigenvectors for a 3 x 3 real symmetric matrix with Numerical Recipes. The answer was not the same as in Schaum's Outline. I think the problem is that eigenvectors corresponding to degenerate eigenvalues (multiplicity > 1) are not necessarily mutually orthogonal, though they must satisfy the matrix eigenvalue equation. The routines in Numerical Recipes make them orthogonal. As with Schaum’s Outline, Matlab may not make them mutually orthogonal.

Please see the thread:

evecs not correct with jacobi, tred2, tqli in this forum.

Ward Ciac