Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

enter image description here

as $A$ is hermitian positive definite so all eigen values are real and strictly positive. we can diagonalize $A=PDP^{-1}$ for some invertible $P$ and we can chose $B=P\sqrt{D}P^{-1}$ and hence $a$ is true?

$BB'$ is symmetric that is true, I am not sure about positive definiteness, let $X$ be any eigen vector corresponding to the eigen value $\lambda$ then $BB^TX=\lambda X\Rightarrow XBB^TX=\lambda XX^T\Rightarrow X(BB^T-\lambda I)X^T=0$.well please help.

share|cite|improve this question
For (a), what can you say about the eigenvalues of a positive definite Hermitian matrix? Can you take their square roots? For (b), note that $\langle x , B^*Bx \rangle = \|Bx\|^2$. $B$ is non-singular, what does that tell you about $\|Bx\|$? – copper.hat Dec 17 '12 at 18:03
I do not get how $<x,B^* Bx>=< B^*x,Bx> =||Bx||^2$ and I can not getting any clue as $B$ is any arbitrary matrix. – Un Chien Andalou Dec 17 '12 at 18:17
$B$ is not arbitrary, it is non-singular. A basic (defining, actually) property of adjoints is $\langle x, A^* y \rangle = \langle A x, y \rangle$. – copper.hat Dec 17 '12 at 18:19
well $BB^T$ need not be positive definite always i guess, I got one counter example of a symmetric matrix which is not positive definite. – Un Chien Andalou Dec 17 '12 at 18:40
$\langle x , B^*Bx \rangle = \langle Bx , Bx \rangle = \|Bx\|^2$, and since $B$ is invertible, you have $\|Bx\|=0$ iff $x=0$. What does that say about $B^*B$. – copper.hat Dec 17 '12 at 19:49
up vote 0 down vote accepted

(a) You wrote $A=PDP^{-1}$ for some invertible $P$ and chose $B=P\sqrt{D}P^{-1}$. This is not good enough. We need $B$ to be positive definite, but your $B$ is not necessarily Hermitian. But you are close. Note that Hermitian matrices can be diagonalized in a very special way (which is described in Wikipedia).

(b) You need to check that $x^TB^TBx>0$ for every real vector $x$ (that's what "positive definite" means). Note that $x^TB^TBx=(Bx)^T(Bx)$.

share|cite|improve this answer
I have checked that if $B$ is invertible then the assertion is true, so both the statement $a,b$ are true? right? – Un Chien Andalou Jun 17 '13 at 14:25
@TaxiDriver Yes, both statements are true, but I don't get the "so" in your comment. That $B$ is invertible in (b) has nothing to do with statement (a). How did you conclude that "so statement a is true"? – user1551 Jun 17 '13 at 15:06
No No, "so" was just a use to make sure whether both are true or not – Un Chien Andalou Jun 17 '13 at 15:09

Your Answer


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.