Given $\Omega\subset \mathbb R^N$ is open and let function $\varphi$: $\Omega\to [1,+\infty]$, $\varphi\in L^1_{loc}(\Omega)$ be given. Suppose $\mu$ is a finite Radon measure on $\Omega$ and we future have $$ \int_\Omega \varphi \,d\mu<\infty. $$ Then can I claim that $\varphi\mu$ is also a finite Radon measure? It looks true but I would like to have some details.

Thank you!

My try:

First of all, for each compact set $K\subset \Omega$, we have $$ \int_K \varphi\,d\mu\leq \int_\Omega \varphi\,d\mu<\infty. $$ Now I only need to show $\varphi\mu$ is Borel regular. Let $\nu:=\varphi \mu$, we have $\nu$ is Borel since $\varphi$ is Borel measurable and $\mu$ is Radon.

To finish my proof, I need to show for each $A\subset \Omega$, there exists a Borel set $B$ such that $A\subset B$ and $\nu(A)=\nu(B)$, and I got stuck here...

I found another similar post here... Please feel free to close mine.

  • $\begingroup$ Why do you think $\int_\Omega \varphi \, d\mu < \infty$? You only have $\varphi \in L_{loc}^1(\Omega)$, not $\varphi \in L^1(\Omega)$. Furthermore, I don't think your definition of "Radon measure" is the standard one. Are you reading Evans & Gariepy? $\endgroup$ – PhoemueX Aug 7 '15 at 18:17
  • $\begingroup$ @PhoemueX Yes I am reading Evans & Gariepy. The part $\int_\Omega \varphi \,d\mu<\infty$ is the assumption. $\endgroup$ – spatially Aug 7 '15 at 18:19

Disclaimer: This answer may be slightly off, since I am not that familiar anymore with the nonstandard definitions of Evans & Gariepy. Anyway...

Note that $\infty > \int_\Omega \varphi \, d\mu \geq \int_\Omega d\mu = \mu(\Omega)$.

Let $A \subset \Bbb{R}^N$. Since $\mu$ is Radon, there is $B \supset A \cap \Omega$ with $\mu(B) = \mu(A \cap \Omega)$. We can assume $B\subset \Omega$ (why?).

Since $B$ is Borel and hence $\mu$-measurable and of finite measure, we get

$$ \mu(B \setminus (A \cap \Omega)) = \mu(B) - \mu(A\cap \Omega) = 0 $$ and thus

$$ \nu((B \cup \Omega^c) \setminus A) = \int_{((B \cup \Omega^c) \setminus A) \cap \Omega} \varphi \, dx = \int_{B \setminus (A\cap \Omega)} \varphi \, dx = 0, $$ which easily yields $$ \nu(A) \leq \nu(B \cup \Omega^c) \leq \nu((B \cup \Omega^c) \setminus A) + \nu(A) = \nu(A), $$ so that $B \cup \Omega^c$ is a Borel set with $B \cup \Omega^c \supset A$ and $\nu(A) = \nu(B \cup \Omega^c)$.


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.