Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $x$ and $y$ be unit vectors in the Euclidean norm. Define $s(\lambda) = |y^Hx|$, where $Ax=\lambda x$ and $ y^HA=\lambda y^H$. Here $\lambda$ is a simple eigenvalue (an eigenvalue with algebraic multiplicity $1$). I would like to prove $s(\lambda) \leq 1$ and $s(\lambda)\neq 0$. I tried the following and gave up after that.

$$Ax=\lambda x $$
Premultiply by $y^H$
$$\begin{align*} y^HAx &= \lambda y^Hx \\ y^HAx &= \lambda y^H I x \end{align*}$$
Taking norm on both sides,
$$ \begin{align*} \frac{1}{\lambda} \|y^HAx\| &= \|y^H I x \| \\ \frac{1}{\lambda} \|y^HAx\| &= |y^Hx| \end{align*}$$

Can anyone help me how to proceed after this? (Similar question, with a matrix free proof here:Proving that two right and left eigenvectors are not orthogonal.)

share|improve this question
1  
$s(\lambda)\le1$ is immediate from the fact that $x$ and $y$ are unit vectors; it has nothing to do with them being eigenvectors of $A$. So you only need to show $s(\lambda)\neq0$. –  joriki Nov 2 '11 at 7:52

1 Answer 1

up vote 1 down vote accepted

In your attempt, you wrote $\frac1\lambda$, which is not a valid fraction if $\lambda=0$. Also, you haven't made use of the fact that $\lambda$ is a simple eigenvalue. Perhaps you should try another approach. For instance, assume that $x$ is a unit vector. If $U$ is a unitary matrix with $x$ as its first column (i.e. $U(1,0,\ldots,0)^\top = x$), then $$ A = U \begin{pmatrix} \lambda&v^\top\\ 0&B \end{pmatrix} U^\ast = U\widehat{A}U^\ast \textrm{ (say)} $$ for some vector $v$ and some matrix $B$. Since $U$ is unitary, eigenvalues, algebraic/geometric multiplicities of eigenvalues as well as orthogonality of vectors are preserved. Yet it is easier to make your argument by considering $\widehat{A}$. For example, that $\lambda$ is a simple eigenvalue means $B-\lambda I$ is invertible.

share|improve this answer

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.