1
$\begingroup$

Here is a lemma that I must prove as part of my midterm exam tomorrow:

Suppose $R\subset \mathbb{R}^n$ is a rectangle, $U\supset \overline R$, and $N\subset \overline R$ has measure zero. Let $\phi: U\to > \mathbb{R}^n$ be $C^1(U)$ i.e. its first derivative is continuous. Then $\phi(N)$ has measure zero.

Here is the proof that our teacher provided us in class. I bolded the part that I disagree with/do not understand.

$\phi \in C^1(U)$, $\overline R$ is compact, so $D\phi$ is bounded, i.e. $|D\phi (x)|\le M\ \forall x\in \overline R$.

So, $\forall x,y\in \overline R$, $\| \phi (x)-\phi (y)\| \le M*\| x-y\|$

Given $\epsilon >0$, let ${R_i}$ be a countable set of rectangles that cover N, with $\sum_{i=1}^{\infty}v(R_i)<\epsilon$.

If $x,y\in R_i$, then $\phi(x)$ and $\phi(y)$ are contained in a rectangle $R_i'$ of diameter $M*diam(R_i)$.

The volume of this rectangle is $(M*diam(R_i))^n\le M^n*v(R_i)$, so $\phi (N)$ is covered by rectangles $R_i'$ s.t. $\sum_{i=1}^{\infty}v(R_i')\le > \sum_{i=1}^{\infty}M^n*v(R_i)<M^n*\epsilon$.

Our given definition of the volume of a rectangle $R\subset \mathbb{R}^n$, with $R=[a_1,b_1]\times ...\times [a_n,b_n]$ is $\prod_{i=1}^n(b_i-a_i)$.

Our given definition of the diameter of a rectangle is $\sup_{x,y\in R}{\|y-x\|}$.

As far as I can tell, the volume of $R_i'$ should be $\le (M*diam(R_i))^n$, and $(M*diam(R_i))^n$ should be $\ge M^n*v(R_i)$. What's gone wrong? Can someone provide an alternative proof? It's getting really late and I need to get to bed, so I would really appreciate help ASAP. Thanks!

Edit: I think I might have come a solution:

$v(R_i)=\prod (a_i-b_i)$.

Thus $v(R_i')\le \prod (\phi (a_i)-\phi (b_i))\le \prod M*(a_i-b_i)=M^n*v(R_i)<M^n*\epsilon$

$\endgroup$
2
$\begingroup$

Thanks for explaining what you have tried! I agree with you in that I think there is an error here. What the bound should be, I believe, is $(M\cdot \text{diam}(R_i))^n\leq CM^n\cdot v(R_i)$ where $C$ is some constant independent of $R_i$ (i.e., independent of the rectangle).

In other words, we would like to find a constant $C$ (independent of $R$ in the following inequality) such that $(\text{diam}(R))^n\leq Cv(R)$ for all rectangles $R$. Can you think of what $C$ should be? It could possibly depend on the dimension, $n$.

Hope this helps! Please let me know if you would like to discuss this further.

$\endgroup$
0

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.