3
$\begingroup$

$\newcommand{\spec}{\operatorname{Spec}}\spec\sqrt2=\{\lfloor k\sqrt2\rfloor: k \ge 0\}$.

I have no idea of how I can prove the statement in the question.

Prove that $\spec\sqrt2$ contains infinitely many powers of $2$.

$\endgroup$
12
  • 2
    $\begingroup$ The most obvious place to start would be to enumerate the first few values of $\lfloor k \cdot \sqrt{2}\rfloor$ to get a feel for how it behaves. The next most obvious is to devise an algorithm for solving $\lfloor k \cdot \sqrt{2} \rfloor = 2^n$ for $k$. $\endgroup$
    – user14972
    Commented Oct 5, 2012 at 21:10
  • 1
    $\begingroup$ What a strange notation. Why $\mathrm{Spec} \sqrt{2}$? $\endgroup$ Commented Oct 5, 2012 at 21:12
  • $\begingroup$ @Alexander: Graham, Knuth, & Patashnik, Concrete Mathematics, refer to this sequence as the spectrum of $\sqrt2$. $\endgroup$ Commented Oct 5, 2012 at 21:12
  • $\begingroup$ As stated, this question is obviously true. Do you mean infinitely many integer powers of 2? $\endgroup$
    – GeoffDS
    Commented Oct 5, 2012 at 21:14
  • $\begingroup$ @Graphth: Of course. $\endgroup$ Commented Oct 5, 2012 at 21:15

1 Answer 1

2
$\begingroup$

Let $k=\lceil 2^n\sqrt 2\rceil$. Then $2^n\sqrt 2<k<2^n\sqrt 2+1$. In fact we have either $2^n\sqrt 2<k<2^n\sqrt 2+\frac12$ or $2^n\sqrt 2+\frac12<k<2^n\sqrt 2+1$, depending on the ($n+1)$st binary digit of $\sqrt 2$ (which becomes the first digit of $2^n\sqrt 2$). Since $\sqrt 2$ is irrational, there are infinitely many $n$ (and correspondingly infinitely many $k$) such that $$2^n\sqrt 2<k<2^n\sqrt 2+\frac12$$ holds. Together with $k\sqrt 2 -1<\lfloor k\sqrt 2\rfloor <k\sqrt 2$ we find $$ 2^n\cdot 2-1 <\lfloor k\sqrt 2\rfloor <2^n\cdot 2+\frac{\sqrt 2}2,$$ hence $\lfloor k\sqrt 2\rfloor=2^{n+1}$.

$\endgroup$
1
  • $\begingroup$ Thanks for the answer, but I see a lot of unexplained things, which I don't seem to be able to comprehend :( . (n+1)st binary digit is the one counted from the left end, right? And, how does this bit influence the range of k like you posted? Infinitely many k? But k is a natural number, right? How would k take infinitely many values in between 2 real numbers differing by 1/2? $\endgroup$
    – learner
    Commented Oct 5, 2012 at 21:55

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .