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.

Can somebody please help me understand this Cauchy-Schwarz inequality? I am told that $A\neq 0$ is Hermitian with nonzero eigenvalues $\lambda_{1},\cdots,\lambda_{r}$. And according to the Cauchy-Schwarz inequality it is obvious that...

$\newcommand{\tr}{\mathrm{tr}}(\tr (A))^{2}=(\Sigma_{i=1}^{k}(\lambda_{i}))^{2}\leq k\Sigma_{i=1}^{k}\lambda_{i}^{2}=k (\tr A^{2}) $

However, I do not follow this reasoning. Thanks in advance.

share|improve this question
    
Which part don't you understand? It must be the left-hand equality and/or the right-hand one. (Hint: Spectral theorem for Hermition matrices for LHS. $A^2 = A A = A^\ast A$ for RHS.) –  cardinal Nov 11 '11 at 1:21
    
well, for one, the above implies that $0\leq k \Sigma_{i=1}^{k} \lambda_{i}^{2}- (\Sigma_{i=1}^{k}\lambda_{i})^{2}$ but this is not clear to me. Can we rewrite this so that it becomes clear when this inequality equals zero? –  Edison Nov 11 '11 at 2:06

2 Answers 2

up vote 1 down vote accepted

Let $V$ be a vector space over the the complex numbers and let $\langle \cdot , \cdot \rangle : V \times V \rightarrow \mathbb{C}$ be an inner product map. Note that the map $\left\| \cdot \right\|:V \rightarrow \mathbb{R}^+$ given by $v \mapsto \left\|v\right\| = \langle v,v \rangle^{1/2}$ is a norm map. Then the Cauchy-Schwarz inequality says that for any $u, v \in V$ we have that $\left|\langle u,v \rangle\right| \le \left\|u\right\| \left\|v\right\|$ or $\langle u,v \rangle^2 \le \langle u,u \rangle \langle v,v \rangle$. Let $M_k(\mathbb{C})$ denote the $k \times k$ matrices over the complex numbers. Then the map $M_k(\mathbb{C}) \times M_k(\mathbb{C}) \rightarrow \mathbb{C}$ given by $(A,B) \mapsto Tr(AB^*)$ is an inner product (verify this) and its induced norm is the Frobenius norm. The Cauchy-Schwarz inequality becomes $Tr(AB^*)^2 \le Tr(AA^*) \cdot Tr(BB^*)$. Now take $A$ to be Hermitian and $B$ to be the identity. This gives $\left(Tr(A)\right)^2 \le k Tr(A^2)$, since $Tr(I_k I_k^*) = Tr(I_k) = k$.

share|improve this answer

C-S inequality says $$|x_1.y_1+x_2.y_2+\cdots+x_r.y_r|^2\le|x_1^2+x_2^2+\cdots+x_r^2|.|y_1^2+y_2^2+\cdots+y_r^2|$$

Now put $x_i=1$ and $y_i=\lambda_i$.

share|improve this answer
    
Equality holds iff $x$ and $y$ are dependent i.e., either is a null vector or $x=cy$. So, in your case iff $\lambda=(c,c,\cdots,c)$. –  Tapu Nov 11 '11 at 2:24

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.