Exercise $12.10$ - Francesco Maggi's book

78 Views Asked by At

$\textbf{Exercise 12.10}$ If $E$ and $F$ are sets of locally finite perimeter, then $\mu_E = \mu_F$ on $\mathcal{B}_b(\mathbb{R}^n)$ if and only if $|E \Delta F| = 0$. Characterize the case $\mu_E = - \mu_F$ on $\mathcal{B}_b(\mathbb{R}^n)$.

I know that

Let $E$ a Lebesgue measurable set in $\mathbb{R}^n$. We say that $E$ is a $\textbf{set of locally finite perimeter}$ in $\mathbb{R}^n$ if for every compact set $K \subset \mathbb{R}^n$ we have

$$\sup \left\{ \int_E \text{div} T(x) dx \ ; \ T \in C_c^1(\mathbb{R}^n,\mathbb{R}^n), \text{spt} K, \sup_\limits{\mathbb{R}^n} |T| \leq 1 \right\} < \infty. \ (12.1)$$

If this quantity is bounded independently of $K$, then we say that $E$ is a $\textbf{set of finite perimeter}$ in $\mathbb{R}^n$.

$\textbf{Proposition 12.1}$ If $E$ is a Lebesgue measurable set in $\mathbb{R}^n$, then $E$ is a set of locally finite perimeter if and only if there exists a $\mathbb{R}^n$-valued Radon measure on $\mathbb{R}^n$ such that

$$\int_E \text{div} T = \int_{\mathbb{R}^n} T \cdot d\mu_E, \forall T \in C_c^1(\mathbb{R}^n,\mathbb{R}^n). \ (12.2)$$

Moreover, $E$ is a set of finite perimeter if and only if $|\mu_E|(\mathbb{R}^n) < \infty$.

$\textbf{Remark 12.2}$ Of course, $(12.2)$ is equivalent to

$$\int_E \nabla \varphi = \int_{\mathbb{R}^n} \varphi d\mu_E, \forall \varphi \in C_c^1(\mathbb{R}^n).$$

I would like to know if what I did is right.

$(\Longrightarrow)$

If $\mu_E = \mu_F$ on $\mathcal{B}_b(\mathbb{R}^n)$, then

$$\int_E \nabla \varphi = \int_{\mathbb{R}^n} \varphi d\mu_E = \int_F \nabla \varphi, \forall \varphi \in C_c^1(\mathbb{R}^n).$$

by the remark $12.2$. Thus, $\int_{\mathbb{R}^n} \nabla \varphi (\chi_E - \chi_F) = 0$, which imply that $\chi_E - \chi_F = 0$ at $|\cdot|$-a.e. (I'm denoting by $\mu$ the measures $\mu_E$ and $\mu_F$ because these measures are equal by hypothesis and $|\cdot|$ denotes the Lebesgue measure on $\mathbb{R}^n$), but $\chi_E \neq \chi_F$ on $E \Delta F$, then $|E \Delta F| = 0$.

$(\Longleftarrow)$

If $|E \Delta F| = 0$, then $|E \backslash F| = |F \backslash E| = 0$. I want to use this to state

$$\int_E \nabla \varphi = \int_{E \backslash F} \nabla \varphi + \int_{E \cap F} \nabla \varphi = \int_{E \cap F} \nabla \varphi = \int_{F \backslash E} \nabla \varphi + \int_{E \cap F} \nabla \varphi = \int_F \nabla \varphi,$$

but this imply that $\int_{E \backslash F} \nabla \varphi = \int_{F \backslash E} \nabla \varphi = 0$?

I want use this, the hypothesis that $E$ and $F$ are sets of locally finite perimeter and the remark $12.2$ to ensure that $\mu_E = \mu_F$ on $\mathcal{B}_b(\mathbb{R}^n)$, but why the measures agree only on $\mathcal{B}_b(\mathbb{R}^n)$?

$\textbf{P.S.:}$ I believe $\mathcal{B}_b(\mathbb{R}^n)$ denotes the family of the bounded Borel sets of $\mathbb{R}^n$, because the family of the Borel sets of $\mathbb{R}^n$ is denoted by $\mathcal{B}(\mathbb{R}^n)$.