3
$\begingroup$

A matrix $A$ is normal if $AA^*=A^*A$. Suppose $(\lambda,x)$ is an eigenpair of $A$, i.e., $Ax = \lambda x$. Proof for a normal matrix $A$, $(\lambda,x)$ is an eigenpair of $A$ if and only if $(\bar{\lambda},x)$ is an eigenpair of $A^*$.

My try:

We have $Ax = \lambda x$ and want to show $A^*x=\bar{\lambda}x$. Taking the conjugate transpose we have $x^*A^* = \bar{\lambda} x^*$. Now we can write $x^*A^* A= \bar{\lambda} x^*A$, using the fact that $A$ is normal, we have $x^*A A^*= \bar{\lambda} x^*A$. Now I do not know how to get $A^*x=\bar{\lambda}x$.

$\endgroup$
1
$\begingroup$

Proving one direction is enough. Think in norm. Equality in normed space means zero in norm: $\Vert (A-\lambda I) x \Vert = 0$, then take transpose inside: $\Vert x^* (A^*- \lambda I) \Vert = 0$. Square both sides and appy the given condition that $A$ is normal to get $$ x^* AA^* x +\Vert \lambda\Vert^2 x^* x -\bar{\lambda} x^* A x - \lambda x^* A^* x.$$ Factorize this into $\Vert (A^*-I) x \Vert^2 = 0$.

$\endgroup$

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.