MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

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

enter image description here $\mu_{n-1}$ is the $n-1$ dimensional measure and $\operatorname{meas}$ is the $n$-dimensional one.

$I(\varrho)$ is the ball of radius $\varrho$ around a fixed point $y$ in the domain $\Omega\subset \mathbb{R}^n$ and $A(t,\varrho)$is the subset of $I(\varrho)$ where a certain function $u$ is greater than $t$.

Surely $I(\varrho) \geq 2\tau(t;\varrho)$ infact, $$\operatorname{meas}{A(t;\varrho)}+\operatorname{meas}(I(\varrho)\setminus A(t;\varrho))= \operatorname{meas}I(\varrho)$$

It seems to me that it is claimed $x > y $ implies $ A-x > A-y$

Fortunately, it turns out the inequality still holds but you need a bit of work (EDIT)

Original article: De Giorgi's original article pages 5-6


Let us assume that $\operatorname{meas}(I(\varrho))=1$ then $\tau(t;\varrho)=x\leq \frac{1}{2}$ what we need to prove is that for $\alpha < 1$ $$(1-x)^{\alpha} + x^{\alpha} -1 \geq 2x^{\alpha}-(2x)^{\alpha}$$

but considering $$f(x)=\left( \frac{1-x}{x}\right) ^{\alpha} -\frac{1}{x^{\alpha}}$$

we can prove that it is decreasing in $(0,\frac{1}{2})$ and $f(\frac{1}{2})=1-2^{\alpha}$

From this we obtain the inequality and hence justify the line.

Due to the several mistakes the question has changed a lot since I posted it, just to be clear I am asking the following now:

1) Is my "fixing" of the proof alright or am I oversimplifying things?

2) Do you think it was an actual mistake that fortunately did not lead to a problem or it is something obvious enough to be omitted in an article?

P.S. sorry if the question is not specific enough, I think it was when I first posted it!

share|cite|improve this question
up vote 5 down vote accepted

Your explanation looks correct, but there are other ways to see it. For example, it's a special case of Karamata's majorization inequality for the concave function $f(z) := z^\alpha$. When $0 \le x \le 1/2$, the sequence $1,x$ majorizes $1-x,2x$, so $f(1) + f(x) \le f(1-x) + f(2x)$, which is the inequality you want. (And Karamata's inequality is overkill, since the version for sequences of length 2 is easier to prove. All you need to check is that $f(z)+f(y-z)$ is an increasing function of $z$ for $z \le y/2$, which follows immediately from concavity. Then apply this with $y=1+x$.)

I don't see any reason to think De Giorgi made a mistake here, and I would assume he had an argument based on concavity. Maybe there's a one-line version he thought was obvious.

However, I wouldn't say it's obvious the way he has written it, and he could have made it easier to read with a little more explanation. Even a hint like "and it follows from the concavity of $z \mapsto z^{1-1/r}$ that..." could help.

share|cite|improve this answer
thank you, I thought it could have been a mistake since the rest of the paper explains quite carefully every step (for istance, the previous inequality $meas(I(\varrho)) \geq 2\tau$ is more trivial than this!) possibly he just overlooked it, or maybe you are right and there is some trivial justification! – Moritzplatz Jun 30 '13 at 18:40

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.