I needed to find a lower bound on $|\det(A+B)|$ where $|.|$ is the absolute value operator.

Because I was unable to get such a bound so I was trying to guess a bound and prove it.

But $||\det(A)|-|\det(B)||$ does not seem to work.

If $\sigma_1$, $\sigma_2$, ..., $\sigma_n$ are singular values of $A$, and $\mu_1$, $\mu_2$, ..., $\mu_n$ are singular values of $B$, then $|(\sigma_1-\mu_1)(\sigma_2-\mu_2)\cdots(\sigma_n-\mu_n)|$ does not seem to work as the lower bound either.

By "does not seem to work" I mean it is not the lower bound of $|\det(A+B)|$.

Any pointers on how to proceed will be helpful. A bound in case there is a significant difference between the singular values of $A$ and $B$ may also work.

The only case that I can solve for is when both $A$ and $B$ are positive semi definite, where $|\det(A+B)|=\det(A+B)\geq \det(A)+\det(B)=|\det(A)|+|\det(B)|.$


The trivial solution would be: the lower bound of $|\det(A+B)|$ is $0$. Here is a bit more general solution: under some conditions $\det(A+B) \geq \det(A) + \det(B)$ (see https://mathoverflow.net/questions/65424/determinant-of-sum-of-positive-definite-matrices). If $\det(A+B) \geq 0$, this means $|\det(A+B)| = \det(A+B) \geq \det(A) + \det(B)$. This solution is more informative then the trivial one (as longs as $\det(A) + \det(B) >0$). The Wikipedia page on the Matrix determinant lemma might help aswell.

  • $\begingroup$ Thanks, the inequality you say only works for positive semi definite matrices. I think I was aware of that when I posted the question 11 months ago. The question is intended for the general case. $\endgroup$ – jinkx Jun 3 '16 at 7:36
  • $\begingroup$ OK, then it might help if you add to your question which cases you allready considered and what did you achive/where do you get stuck. $\endgroup$ – Qaswed Jun 6 '16 at 8:16

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.