An application of Lebesgue Differentiation Theorem

507 Views Asked by At

Let $\delta_0$ be the measure defined as $\delta_0(A)=1$ if $0 \in A$ and $\delta_0(A)=0$. Let $g_h(x)=\frac{1}{h}1_{[-h,h]}$ and define $\nu_h$ as:

$\nu_h(A)=\int_{A}g_h(x)dx$.

Show that for every $f$ continuous we have

$lim_{h \to 0}\int_{\mathbb{R}}fd\nu_h=\int_{\mathbb{R}}fd\delta_0$.

My attempt:

So, I think it's a tricky application of Lebesgue Differentiation Theorem which states that(as in folland): If $k\in L^1$ then for almost every x:

$lim_{r \to 0} \frac{1}{m(E_r)}\int_{E_r}k(y)dy=k(x)$

for every family $E_r$ that shrink nicely to x.

So for us, $E_r=[-h,h]$, then

$\nu_{h}(A)=\int_{E_r}g_h(x)dx=\int_{E_r}\frac{1}{2h}1_{[-h,h]}dx=\frac{1}{2h}\mu(E_r\cap[-h,h])=\frac{2h}{2h}=1$,

and then I don't know how to get the function $k$ to apply LDT.

Thanks for any help!

1

There are 1 best solutions below

0
On BEST ANSWER

For each $f\in L^1(\mathbb R)$, we have $\int fd\nu_h=\int fg_hdx=\frac{1}{h}\int_{-h}^{h}fdx.$

Applying the Lebesgue Density Theorem, we get

$\lim_{h\to 0}\int fdv_h=f(0).$

On the other hand, it follows from the definition of $\delta_0$ that

$\int fd\delta_0=\int\chi_{(-1/n,1/n)}\cdot fd\delta_0$

and since $\chi_{(-1/n,1/n)}\cdot f\le |f|,$

the Dominated Convergence Theorem gives

$\int fd\delta_0=\lim\int fd\delta_0=\lim\int\chi_{(-1/n,1/n)}\cdot fd\delta_0=\int\lim\chi_{(-1/n,1/n)}\cdot fd\delta_0=f(0)\cdot \delta_0(0)=f(0)$

and we conclude that $\lim_{h\to 0}\int fdv_h=\int fd\delta_0.$