1
$\begingroup$

Suppose $n$ is a positive integer, $x_i \geq 0$ and $\alpha >\beta >0$, is the following inequality true ?
$\left ( \frac{\left ( \sum_{i=1}^{n} x_{i}^{\alpha} \right )^2-\left ( n-1 \right )\sum_{i=1}^{n} x_{i}^{2\alpha}}{n} \right)^{\frac{1}{\alpha}}\leq \left ( \frac{\left ( \sum_{i=1}^{n} x_{i}^{\beta} \right )^2-\left ( n-1 \right )\sum_{i=1}^{n} x_{i}^{2\beta}}{n} \right)^{\frac{1}{\beta}}$
Where we should assume $\left ( \sum_{i=1}^{n} x_{i}^{\alpha} \right )^2-\left ( n-1 \right )\sum_{i=1}^{n} x_{i}^{2\alpha}\geq 0$ and $\left ( \sum_{i=1}^{n} x_{i}^{\beta} \right )^2-\left ( n-1 \right )\sum_{i=1}^{n} x_{i}^{2\beta}\geq 0$
The case for $n=1,2$ is obvious, other cases remain unsolved.
(BTW, try not to use EV theorem to solve this problem)
If you also find it hard to tackle, what if $\alpha=2$ and $\beta=1$ ?

$\endgroup$
4
  • $\begingroup$ Why "try not to use EV theorem to solve this problem"? $\endgroup$
    – LSpice
    Commented Oct 27 at 23:36
  • $\begingroup$ Actually, EV theorem is almost useless in this inequality... $\endgroup$ Commented Oct 30 at 14:41
  • $\begingroup$ Re, ah, so this is just a helpful suggestion to avoid a dead end rather than a request for a novel proof technique? $\endgroup$
    – LSpice
    Commented Oct 30 at 16:13
  • $\begingroup$ As far as I concerned, yes $\endgroup$ Commented Oct 30 at 16:23

0

You must log in to answer this question.

Browse other questions tagged .