3
$\begingroup$

see:old post and Make the integral discrete, we have

conjecture :for any complex numbers $a_{i},i=1,2,\cdots,n$,and $p_{i}\ge 0$ such $p_{1}+p_{2}+\cdots+p_{n}=1$,then maybe be have $$\sum_{i=1}^{n}\sum_{j=1}^{n}|a_{i}+a_{j}|\cdot p_{i}p_{j}\ge \sum_{i=1}^{n}p_{i}|a_{i}|$$

How prove this complex inequality?Thanks

take $p_{i}=\dfrac{1}{n},i=1,2,\cdots,n$,and Amuess that $a_{i}\in R$.I have prove that following

**Theorem **.$$\sum_{i,j=1}^{n}|a_{i}+a_{j}|\ge n\sum_{i=1}^{n}|a_{i}|$$ proof: Equality occurs if all $x_k$ are null, or else if $n$ is even, half the variables are equal to any real number $a$, and the other half are equal to $-a$. The proof goes along the lines exposed in the above, but an alternate solution, given by C. Tanasescu, is worth mentioning.

Use the obvious equality $|a| + |b| - |a+b| = 0$ if $ab\geq 0$ and $= 2\min\{|a|,|b|\}$ if $ab<0$. Then $$\begin{align*}\sum_{1\leq i,j\leq n} (|x_i| + |x_j| - |x_i+x_j|) &= \sum_{x_ix_j< 0} 2\min\{|x_i|,|x_j|\}\\ & = 4\sum_{x_i>0>x_j}\min\{x_i,-x_j\} \leq 4\sum_{x_i>0>x_j}\sqrt{-x_ix_j} \\ &= 4\sum_{x_i>0}\sqrt{x_i} \sum_{0>x_j}\sqrt{-x_j} \leq \left (\sum_{x_i>0}\sqrt{x_i} + \sum_{0>x_j}\sqrt{-x_j}\right )^2 \\ &=\left (\sum_{1\leq k\leq n}\sqrt{|x_k|}\right )^2 \\ &\leq n\sum_{1\leq k\leq n}|x_k|\end{align*}$$

It follows $\sum_{1\leq i,j\leq n} |x_i+x_j| + n\sum_{1\leq k\leq n}|x_k| \geq \sum_{1\leq i,j\leq n} (|x_i| + |x_j|) = 2n\sum_{1\leq k\leq n}|x_k|$, which is equivalent to the required inequality

$\endgroup$
7
  • 7
    $\begingroup$ All your conjectures here have been settled in the "old post" thread, so I see no reason to open a new one. Besides, it is fairly well known that $4=2\times 2$, so the $2$ function inequality immediately implies the $4$ function one... $\endgroup$
    – fedja
    Commented Feb 6, 2021 at 2:44
  • 2
    $\begingroup$ @fedja: Maybe this post was edited, but I think the point of the new question is that the variables are now complex numbers. $\endgroup$
    – GH from MO
    Commented Feb 9, 2021 at 10:10
  • 3
    $\begingroup$ @GHfromMO Come on! Just project and average, that's it :-) $\endgroup$
    – fedja
    Commented Feb 12, 2021 at 0:29
  • 1
    $\begingroup$ @fedja: So the inequality also works in $\mathbb{R}^N$ (hence also in any Hilbert space), right? $\endgroup$
    – GH from MO
    Commented Feb 12, 2021 at 1:18
  • 1
    $\begingroup$ @GHfromMO Yes, of course. Now, if you want values in an arbitrary Banach space, it gets trickier and I'm not sure anymore, but that is certainly not what the OP asked. Want to open a separate thread for that? I'm afraid that there may be trivial counterexamples though. $\endgroup$
    – fedja
    Commented Feb 12, 2021 at 1:22

0

You must log in to answer this question.

Browse other questions tagged .