1
$\begingroup$

Let $f: \mathbb{R}^n \rightarrow \mathbb{R}^m $ be Lipschitz and $n \geq m$. A version of the coarea formula says:

$$ \int_A g(x) J_m f(x) d \mathcal{L}^n (x) = \int_{\mathbb{R}^m } \int_{ A \cap f^{-1}(t) } g(p) d \mathcal{H}^{n-m}(p) d \mathcal{L}^m (t) $$

for each Lebesgue measurable subset $A \subset \mathbb{R}^n $ and an integrable function $g: A \rightarrow \mathbb{R} $.

In the proof process, the following map $T$ can be shown to be $\mathcal{L}^m$ measurable:

$$ T(t) = \mathcal{H}^{n-m}( A \cap f^{-1}(t) ) $$

Now consider the case when $A$ is open and the restriction of the map $T$ to $f(A)$: $ T: f(A) \longrightarrow \mathbb{R}^{+} $. I'm wondering if there are extra conditions we can impose on $f$, say it is $C^1$ with Lipschitz Jacobian so that we can estimate $T$'s modulus of continuity (on $f(A)$)?

$\endgroup$

1 Answer 1

2
$\begingroup$

The answer is in the negative. Let $Q=(0,1)\times (0,1)$. Let $I_i$, $i\in\mathbb{N}$ be open intervals in the complement of the ternary Cantor set $C$, and let $$ A=Q\cup\bigcup_{i=1}^\infty I_i\times [1,3/2). $$ That means on top of the square we add rectangles of height $1/2$ with base in the complement of the Cantor set $C$.

Let $f:\mathbb{R}^1\to\mathbb{R}$, $f(x,y)=x$. Then $f(A)=(0,1)$. If $t\in C\cap (0,1)$, then $T(t)=\mathcal{H}^1(f^{-1}(t))=1$ while for $t\in (0,1)\setminus C$, $t\in I_i$ for some $i$ and hence $T(t)=\mathcal{H}^1(f^{-1}(t))=3/2$. Thus the function $T$ is discontinuous at the points of $C$.

Adding more rectangles of height $1/4$ on top of each rectangle $I_i\times [1,3/2)$ and then adding rectangles of height $1/8$ on top of these rectangles and continuing this construction infinitely many times we can produce a bounded open set $A$ such that the function $T$ is discontinuous at almost every point of $(0,1)$.

$\endgroup$
5
  • $\begingroup$ Sorry for the confusion. Your example is not in the intended setting I was thinking about. I have updated the question accordingly. In particular, we consider the restriction of $T$ to $f(A)$. $\endgroup$ Mar 13, 2018 at 0:40
  • $\begingroup$ Does my example answer your question? $\endgroup$ Mar 26, 2018 at 21:02
  • $\begingroup$ Thanks for your answer and sorry for the delayed reply. Your answer helps me think carefully about the needed conditions to well pose the question in my mind. I think your answer is instructive, but I'm not sure if it answers my real question. I will need to think more about it and maybe edit the question accordingly. If you don't mind, please leave your answer here. Thank you again. $\endgroup$ Mar 26, 2018 at 22:22
  • $\begingroup$ If you e-mail me (you [email protected]) I can send you a picture for the domain in my answer so it will help you understand my construction. $\endgroup$ Mar 26, 2018 at 22:26
  • $\begingroup$ It is very kind of you! I kind of get the construction, the question that bothers me now is how much does the continuity of $T$ depend on the domain $A$? If $A$ is a nice shape like a ball or rectangle, can we restrict $f$ so that $T$ is continuous? Obviously, the induced map $T$ depends on both $f$ and $A$. Your construction shows its strong dependence on $A$. $T$'s dependence on $f$ or lack of it remains unclear - which is what I'm initially interested in - but I naively ignored the importance of $A$. $\endgroup$ Mar 27, 2018 at 17:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.