3
$\begingroup$

For positive integer $k$, let $\mu_k=\dfrac{1}{2}\left(\delta(x)+\delta\left(x-\dfrac{2}{3^k}\right)\right)$. Show that $$\lim_{k\rightarrow\infty}(\mu_1\ast\mu_2\ast\cdots\ast\mu_k)((-\infty,x))=C(x),$$ where $C$ is the Cantor function.

The definition of measure convolution that I find is

$$ (\mu_1\ast \mu_2)(I)=\int_\mathbb{R}\int_\mathbb{R}1_I(x_1+x_2)d\mu_1(x_1)d\mu_2(x_2)$$ for any interval $I$.

I'm not sure how to start on this. It's a convolution of $k$ measures...

$\endgroup$
3
$\begingroup$

Here is a probabilistic proof. Recall that a convolution of measures corresponds to taking a sum of independent random variables. Recall that $\mu_k$ is the law of a random variable which takes on value $0$ with probability $\frac{1}{2}$ and $\frac{2}{3^k}$ with probability $\frac{1}{2}$. Therefore, if $X_k$ is sequence of $iid$ random variables taking on the value $2$ with probability $\frac{1}{2}$ and $0$ with probability $\frac{1}{2}$, $\mu_k$ is the law of $\frac{X_k}{3^k}$.

$\mu_1 * \dots * \mu_k$ is therefore the law of $\sum_{i=1}^k \frac{X_i}{3^i}$. It is not hard to see that this sum converges absolutely almost surely (and therefore in distribution) to $\sum_{i=1}^\infty \frac{X_i}{3^i}$ and that the limit random variable is uniformly distributed on the set of numbers which have only $0$ or $2$ in their tenary expansion. This characterizes the limit as the uniform measure on the Cantor set. The Cantor function is the distribution function of that measure.

$\endgroup$

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.