0
$\begingroup$

Prove that, for almost everywhere number $x ∈ [0,1]$ whose decimal expansion contains the block $617$ (for instance, $x = 0.3375617264 ···$), that block occurs infinitely many times in the decimal expansion of $x$. Even more, the block $617$ occurs infinitely many times in the decimal expansion of almost every $x ∈ [0,1]$.

The first part I did taking $E = [0, 617, 0, 618)$ which has positive measure and use that the Lebesgue measure is $f$-invariant where $f(x)=10x- \lfloor 10x \rfloor$ and apply the recurrence theorem of Poincaré.

The second part the book gives a hint:

Note that every interval $J = [ j / 10^k, (j + 1) / 10^k)$ contains subinterval $J'$ such that $\frac{m (J')}{m (J)} = \frac{1}{10^{3}}$ and $f^k (x) ∈ E$ for all $x ∈ J'$. Using Theorem $A.2.14$, conclude that all $x ∈ [0, 1]$ has at least one iterate in $E$. Now the second statement in the exercise the is consequence of the first part.

Theorem $A.2.14.$ Let A be a measurable subset of $R^d$ with Lebesgue measure $m(A)$ positive. Then $m$-almost every $a ∈ A$ is a density point of $A$.

I can not see that there is this interval $ J '$. Someone can explain.

$\endgroup$
  • 1
    $\begingroup$ There's something fishy about the first claim. The number all of whose decimal digits are (say) 3, except for the initial block of 617, viz $x=.61733333...$ looks like a counter example. Maybe you mis-stated what you meant to ask? $\endgroup$ – kimchi lover Mar 13 at 21:43
  • 1
    $\begingroup$ The first claim as stated is plainly wrong. I wonder if what is meant is that it's true of almost every $x \in [0, 1]$. $\endgroup$ – Robert Shore Mar 13 at 21:51
  • 1
    $\begingroup$ The correct is for almost everywhere number $x∈[0,1]$ $\endgroup$ – Ricardo Freire Mar 13 at 22:03
  • $\begingroup$ Borel-Cantelli? $\endgroup$ – Jair Taylor Mar 13 at 22:16
  • $\begingroup$ I can not see relation in using Borel cantelli and explain the existence of the interval $J'$ $\endgroup$ – Ricardo Freire Mar 13 at 22:22
2
$\begingroup$

$$\ J' = \left[\frac{10^3j+617}{10^{k+3}}, \frac{10^3j+618}{10^{k+3}} \right) $$

$\endgroup$

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.