Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I think the following two inequalities are true. However, the proof may not be easy. Does anyone have any hints? Thank you very much!

Fix $a>1$. there exists two constants $K_1$ and $K_2$, such that

$$ \frac{1}{1+|x-y|^{1+a}}\frac{1}{1+|x-z|^{1+a}} \le \frac{K_1}{1+|x-(y+z)/2|^{1+a}} \le \frac{K_2}{1+|x-y|^{(1+a)/2}}\frac{1}{1+|x-z|^{(1+a)/2}} $$

holds for all $x,y,z\in R$.

Thanks a lot!


share|cite|improve this question
Where did these arise? – Alex Becker Feb 24 '12 at 9:31
In my research, I would like to have this inequality. These inequalities are my guess. For sure, I can modify accordingly. It is related to the Poisson kernel and fractional heat equations. – Anand Feb 24 '12 at 11:51
up vote 1 down vote accepted

Let $\alpha=|x-y|$, $\beta=|x-z|$, $\gamma=|x-(y+z)/2|$ and $p=1+a>2$. It is clear that $\gamma\le(\alpha+\beta)/2\le\max(\alpha,\beta)$. Assume without loss of generality that $\alpha\le\beta$. Then $$ \frac1{1+\alpha^p}\,\frac1{1+\beta^p}\le\frac1{1+\beta^p}\le\frac1{1+\gamma^p}, $$ proving the first inequality with constant $K_1=1$. This constant is sharp, as can be seen by letting $\alpha=\beta=\gamma\to0$.

The second inequality reads now as $$ \frac1{1+\gamma^p}\le\frac{C}{(1+\alpha^{p/2})(1+\beta^{p/2})}. $$ or $$ (1+\alpha^{p/2})(1+\beta^{p/2})\le C(1+\gamma^p). $$ This inequality is false. Let $x=(0,\dots,0,1)$, $y=(R,0,\dots,0)$ and $z=(-R,0,\dots,0)$. Then $\alpha=\beta=\sqrt{1+R^2}$ and $\gamma=1$. The left hand side goes t0 $\infty$ as $R\to\infty$, while the right hans side remans bounded.


$\alpha$, $\beta$ and $\gamma$ have a geometrical interpretation. $\alpha$ and $\beta$ are the sides of a triangle, and $\gamma$ is the length of the median from the vertex fo the sides $\alpha$ and $\beta$ to the oposite side.

share|cite|improve this answer
Dear Professor Julián Aguirre, Thank you very much for your solutions. Even we restrict x,y,z to be real numbers, by letting x=1 and y=-z=R, the second inequality will fail as R tends to infinite. Thanks a lot. – Anand Feb 24 '12 at 14:56

Your Answer


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.