Let matrix $ X $ is Hermitian and denote $ \lambda_1(X) \ge \lambda_2(X) \ge \ldots \ge \lambda_n(X) $ eigenvalues of matrix $ X $.

Prove that $ \lambda_i(A + B) \le \lambda_i(A) + \lambda_1(B) $

I think that here could be applied relations that if $ A $ is Hermitian matrix and $ \lambda_1(A) \ge \lambda_2(A) \ge \ldots \ge \lambda_n(A) $ are eigenvalues of $ A $and $ B $ - principal submatrix of order $ n - 1 $ and $ \lambda_1(B) \ge \lambda_2(B) \ge \ldots \ge \lambda_{n-1}(B) $ it's eigenvalues, then $ \lambda_1(A) \ge \lambda_1(B) \ge \lambda_2(A) \ge \ldots \ge \lambda_{n-1}(B) \ge \lambda_n(A) $. Or corollaries from that theorem.

Also I tried to diagonalize matrix using Schur's theorem.

And maybe here could be useful some inequalities connected with eigenvalues, for example:

$ \sum_{i=1}^{k}\lambda_i(A+B) \le \sum_{i=1}^{k}\lambda_i(A) + \sum_{i=1}^{k}\lambda_i(B)$

Or $ \lambda_1(A+B) + \lambda_n(A+B) \le \lambda_1(A) + \lambda_n(A) + \lambda_1(B) + \lambda_2(B)$

But I do not know how to solve my problem. Thanks for the help!


Your Answer

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

Browse other questions tagged or ask your own question.