Is there a known way or software to find integer eigenvectors for an integer matrix with integer eigenvalues?

In particular, I have a large real symmetric matrix with only a small number of distinct eigenvalues. I want to know if it is possible to find an eigenbasis such that each eigenvector contains only -1, 0, and 1 entries.


For an integer matrix $A$ and an integer eigenvalue $\lambda$, $A - \lambda I$ is an integer matrix, and Gaussian elimination will produce rational eigenvectors; multiply by a common denominator and you'll get an integer eigenvector.

But the question in your second paragraph is different: there is an eigenvector with entries $-1, 0, 1$ iff there are two different sets of columns of $A - \lambda I$ with the same sum. Although it's not guaranteed to work for this, I might try to find these using the LLL algorithm.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.