Suppose that $A\in\mathbb{R}^{n\times n}$ is a square matrix. What is the condition for $$(I+\epsilon A)(I+\epsilon A^\top)>\epsilon^2 AA^\top.$$ Here, for matrices $X_1$ and $X_2$, $X_1>X_2$ means that $X_1-X_2$ is positive definite.

Any help would be highly appreciated!


1 Answer 1


Hint: if $B$ is symmetric, then the matrix $I+ \epsilon B$ is positive definite for sufficiently small $\epsilon.$

  • $\begingroup$ But the point is that $A$ is not always symmetric and I want the requirement on $\epsilon$ and the singular values of $A$. $\endgroup$
    – lazyleo
    Jun 28 at 3:18
  • $\begingroup$ @lazyleo You are right! But $AA^T$ is symmetric. $\endgroup$
    – Vezen BU
    Jun 28 at 3:21
  • 3
    $\begingroup$ @VezenBU I think you mean $A + A^T$. $\endgroup$ Jun 28 at 3:37
  • $\begingroup$ @lazyleo Expand the left-hand side, then think about how the hint can be utilised. $\endgroup$ Jun 28 at 6:34

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.