0
$\begingroup$

Why $$\left\|\sum_{i=1}^nx_i\right\|^2\leq n\sum_{i=1}^n\|x_i\|^2$$ for arbitrary norm on the inner product space over the real field? My attempt

$$\left\|\sum_{i=1}^nx_i\right\|^2 = \sum_{i=1}^n\sum_{j=1}^n\langle x_i,x_j\rangle = \sum_{i=1}^n\|x_i\|^2 + \sum_{i=1}^n\sum_{j\ne i}\langle x_i,x_j\rangle \\ \leq \sum_{i=1}^n\|x_i\|^2 + \sum_{i=1}^n\sum_{j>i}\left(\|x_i\|^2 + \|x_j\|^2\right)$$

but then I fail to asses the second term.

$\endgroup$
  • 1
    $\begingroup$ That's not an arbitrary norm, though. How are you defining norm? $\endgroup$ – Thomas Andrews Dec 18 '14 at 15:10
  • 2
    $\begingroup$ Assuming you're willing to use some form of the Cauchy Schwartz inequality, just look at $\sum 1 \cdot x_i$ $\endgroup$ – JHance Dec 18 '14 at 15:11
  • $\begingroup$ @Thomas The norm is the one associated with the inner product $\|x_i\| = \sqrt{\langle x_i,x_i\rangle }$ $\endgroup$ – Aad Dec 18 '14 at 15:38
  • $\begingroup$ @JHance This is very nice. Is it ok, if $x_i$ are, for example, $L^2$ functions? I'm confused by the fact that we define the euclidean norm for vector, whose points are not in $\mathbb{R}$, but functions. $\endgroup$ – Aad Dec 18 '14 at 15:49
1
$\begingroup$

Use this identity

$$\left|\sum_{i=1}^{n}a_{i}b_{i}\right|^2+\sum_{i<j}^{n}\left|a_{i}\overline{b_{j}}-a_{j}\overline{b_{i}}\right|^2=\left(\sum_{i=1}^{n}|a_{i}|^2\right)\left(\sum_{i=1}^{n}|b_{i}|^2\right)$$ let $b_{i}=1$

see:http://en.wikipedia.org/wiki/Lagrange%27s_identity

$\endgroup$

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.