2
$\begingroup$

Suppose I have a tridiagonal square matrix $A$ of some nice form, for which I know the eigenvalues $\lambda_1<\dots<\lambda_n$. $A$ is also essentially nonnegative (nonnegative everywhere except main diagonal).

What can I say about the sum of $A$ and a diagonal matrix $B=\text{diag}(x_1,\dots,x_n)$, for which I know only about sum of it's elements?

Some equalities, maybe inequalities with eigenvalues of the first matrix and values of the second? Special interest is in the maximal eigenvalue

Improve this question
$\endgroup$
4
  • $\begingroup$ I'm not sure I understand the question completely, but the SUM of the diagonal entries certainly can't give anything (in general) on individual eigenvalues ($A=0$). $\endgroup$
    – Christian Remling
    Commented Apr 27, 2014 at 23:38
  • $\begingroup$ assuming a symmetric tridiagonal, isn't the majorization relation $\lambda(A+B)\prec \lambda(A)+\lambda(B)$ sufficient? $\endgroup$
    – Suvrit
    Commented Apr 28, 2014 at 0:34
  • $\begingroup$ @ChristianRemling that is all I know: $A$ is tridiagonal, $B=\text{diag}(x_1,\dots,x_n),\ \sum_{1}^{n} x_i = R$. But any interesting inequality which contain x_i and eigenvalues of $A$ would be a great step for me too $\endgroup$
    – Daniel
    Commented Apr 29, 2014 at 1:09
  • $\begingroup$ @Suvrit I wonder if there is some more interesting property based on tridiagonality and nonnegativity of $A$ and diagonality of $B$ $\endgroup$
    – Daniel
    Commented Apr 29, 2014 at 1:17

0

Reset to default

You must log in to answer this question.

Browse other questions tagged .