-2

Suppose $\lambda$ is the Lebesgue measure on the Borel $\sigma$-algebra.

Does there exist an example of sets $A,B\subset\mathbb{R}$, where:

  1. $A\cup B=\mathbb{R}$
  2. $A\cap B=\emptyset$
  3. for all non-empty open intervals $I:=(a,b)\subset\mathbb{R}$, such that $c\neq 1$ is a non-zero constant:

$$\lambda(A\cap I)=c\cdot\lambda(B \cap I)?$$

This answer might offer a hint, despite $c$ being non-constant, where $c$ is positive (and):

$$\lim_{t\to\infty}\frac{\lambda(A\cap[-t,t])}{2t}=\frac{2}{3}$$

$$\lim_{t\to\infty}\frac{\lambda(B\cap[-t,t])}{2t}=\frac{1}{3}$$

Arbuja
  • 53

1 Answers1

6

$$\def\l{\lambda}$$ Check the Lebesgue Density Theorem. If $A$ is measurable then at almost every point $x\in \mathbb R$ we have that $$\lim_{\epsilon\to0} \frac{\l(A \cap B_{\epsilon}(x))}{\l(B_{\epsilon}(x))} $$ exists and is either equal to $1$ or $0$. This tells us that the only $c$ that can work for your example is $c=0$. Even the value $c=1$ that you want to avoid is impossible (since it would lead to a limiting ration of $1/2$).

Of course $A=\mathbb{Q}$, $B=\mathbb{R}\setminus \mathbb{Q}$ achieves $c=0$.

Jamie Radcliffe
  • 2,259
  • 1
  • 14
  • 9
  • Nevermind. This works when $\varepsilon\to 0$. When $\varepsilon>0$, this is not true. – Arbuja Mar 05 '25 at 01:09
  • 1
    Right, but it does rule out having the ratio be $c$ for all intervals $(a,b)$. – Jamie Radcliffe Mar 05 '25 at 02:56
  • What happens when $c=1$? I was thinking of doing a type of sets $A,B$ such that they divide kind of like cantor but with intervals such that they are measurable. Would it be possible to do that? – MM3 Mar 05 '25 at 10:17