Symmetric matrix with integer values is positive semidefinite

34 Views Asked by At

Disclaimer: This question is a follow-up question to: Symmetric matrix is positive semidefinite. In here, I tried rephrasing the original problem and made some mistakes in the process. This is the correct version of it. Thank you to all that interacted with my previous post.

Let $A=(a_{ij}) \in M_n(\mathbb R)$ be a symmetric matrix such that:

  1. $a_{ij} \in \mathbb N_0, \quad \forall i,j=1,...,n$
  2. $a_{ii} = 2n+1$,
  3. $0 \leq a_{ij} \leq 2n, \quad i \neq j$.

I want to prove that such a matrix is positive semidefinite. It is well known that since $A$ is symmetric, this amounts to show that all eigenvalues of $A$ are non-negative. For that, one has to find a lower bound for the eigenvalues of $A$ that is greater or equal to zero. I tried using Gershgorin theorem, but the bound is not strong enough, the same happens with the one on this paper: https://www.math.uwaterloo.ca/~hwolkowi/henry/reports/bndseigs80.pdf

Does anyone one know stronger bounds on eigenvalues of matrices of this kind? Or to prove directly that this matrix is positive semidefinite?

Thanks in advance!

1

There are 1 best solutions below

1
On BEST ANSWER

Again this is not true.

$$ \det\begin{pmatrix} 9&8&7&6\\ 8&9&0&0\\ 7&0&9&0\\ 6&0&0&9 \end{pmatrix}=-5508. $$ Also, if you replace the $0$s with $1$s the determinant is $-3248$.


Edit: A counterexample when $n=3$: $$ \det \begin{pmatrix} 7&6&5\\ 6&7&1\\ 5&1&7 \end{pmatrix}=-31. $$