I have two sets of numbers satisfying: $x_1+x_2+...+x_n >= y_1+y_2+...+y_n$ Whether it is the case: $x_1^2+x_2^2+...+x_n^2 >= y_1^2+y_2^2+...+y_n^2$ ($x_i^2$ presents the square of xi). $x_i, y_i$ here are not negative numbers.

  • $\begingroup$ Clearly not. Take $0<x_i<-y_i$. $\endgroup$ – Mark Viola Nov 23 '16 at 16:06
  • $\begingroup$ I edited my question. Numbers here are not negative $\endgroup$ – Mr Black Nov 23 '16 at 16:14

This is incorrect. Consider for example $x_1 = \ldots = x_n = 1$, $y_1 = n$ and $y_2 = \ldots y_n = 0$.

Generally speaking, if $x_1 + \ldots + x_n = K$ are nonnegative numbers, then $x_1^2 + \ldots + x_n^2$ can assume any value between $\frac{K^2}{n}$ and $K^2$. The lower bound is attained iff all $x_1$ are equal to $\frac{K}{n}$ and the upper bound is achieved for $x_1 = K$, $x_2 = \ldots = x_n = 0$.


I have a WRONG approach to this problem but I don't know what is wrong, Since

$x_{1}+x_{2}\ldots+x_{n}\ge y_{1}+y_{2}\ldots+y_{n}$

We take integral of both side, due to the Monotony of integral:

$\frac{1}{2}\int(x_{1}+x_{2}\ldots+x_{n})dx \ge \frac{1}{2}\int(y_{1}+y_{2}\ldots+y_{n})dy$

This is a linear combination so:

$\frac{1}{2}\int x_{1}dx_{1}+\ldots +\frac{1}{2}\int x_{n}dx_{n} \ge \frac{1}{2}\int y_{1}dy_{1}+\ldots +\frac{1}{2}\int y_{n}dy_{n}$


$x_{1}^2+x_{2}^2+\ldots+x_{n}^2 \ge y_{1}^2+y_{2}^2+\ldots+y_{n}^2$

What is wrong here?

  • $\begingroup$ Check: You're integrating wrt what over what domain? $\endgroup$ – Macavity Nov 23 '16 at 17:46
  • $\begingroup$ @Macavity I'm integrating wrt to random variables x and y in domain [0,+oo]. If I write this way: $\sum x \ge \sum y => \int \sum x \ge \int \sum y$ $\endgroup$ – Gehua Zhang Nov 23 '16 at 19:43
  • $\begingroup$ And why do you think that would hold true? $\endgroup$ – Macavity Nov 24 '16 at 1:05

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.