Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

I am wondering if there is something known about the interlacing properties of an "Almost Hermitian" matrix, in the following sense: let A be a nxn matrix so that it has a Hermitian principal minor of order (n-1)x(n-1). What interlacing properties does A possess?

Thanks!

share|improve this question
add comment

1 Answer 1

Assuming that the Hermitian minor has distinct eigenvalues, there is nothing we can say about the eigenvalues of $A$. Let the eigenvalues of the Hermitian minor be $\lambda_1$, ..., $\lambda_{n-1}$, so we can choose a matrix where your matrix looks like $$A=\begin{pmatrix} \lambda_1 & & & & a_1 \\ & \lambda_2 & & & a_2 \\ & & \lambda_3 & & a_3 \\ & & & \ddots & & \\ b_1 & b_2 & b_3 & & c \end{pmatrix}$$

The characteristic polynomial of $A$ is $$f(x):=(x-c) \prod_{i} (x-\lambda_i) - \sum_j a_j b_j \prod_{i \neq j} (x-\lambda_i). \quad (\ast)$$ I claim that we can choose $c$ and $a_j b_j$ to make $f(x)$ be any monic degree $n$ polynomial.

Proof: We have $f(\lambda_j) = - a_j b_j \prod_{i \neq j} (\lambda_i - \lambda_j)$. So, assuming that the $\lambda$'s are distinct, we can choose $a_j b_j$ to make $f(\lambda_j)$ have any value. Also, we can use $c$ to fix the value of $f$ at any $x$ other than the $\lambda_i$. A monic polynomial of degree $n$ is determined by its values at $n$ points, so we can arrange for $f$ to be any degree $n$ monic polynomial. $\square$

Equation $(\ast)$ also shows that, if $\lambda_i$ occurs with multiplicity $d$ in the Hermitian minor, then it occurs with multiplicity $\geq d-1$ in $A$.

share|improve this answer
    
Thanks! Follow-up question pops up, though: what if I add the extra constraint that the c and b vectors are multiples of each other? Thanks –  Felix Goldberg Mar 27 '12 at 19:55
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.