3
$\begingroup$

Is it true that $\lim_{x \to \infty} (\left \lfloor{x}\right \rfloor -x) = 0$, or alternatively, $\lim_{x \to \infty} \left \lfloor{x}\right \rfloor=x$? If so, how can we prove it using $\varepsilon$-$\delta$ method?

$\endgroup$
  • 1
    $\begingroup$ No, it's not true. And $\lim_{x\to\infty} \lfloor x\rfloor = x$ doesn't make sense. However, we have $\lim_{x\to\infty} \frac{\lfloor x\rfloor}{x} = 1$. $\endgroup$ – Daniel Fischer Oct 19 '14 at 23:52
  • 1
    $\begingroup$ Take $x_n = n + \frac{1}{2}$ for example, then $x_n \to +\infty$ but $\lfloor{x_n} \rfloor -x_n = -\frac{1}{2}$ $\endgroup$ – Petite Etincelle Oct 19 '14 at 23:54
  • $\begingroup$ I dont see how the two statements are equivalent. Also, I don't see how the limit is zero. $\endgroup$ – FormerMath Oct 19 '14 at 23:55
  • $\begingroup$ I understand that it is not true. But my question actually came from solving a problem in probability theory, which my assumption was the floor(X) converges to X "almost surely." Can I argue the convergence if we are talking about "almost surely"?? $\endgroup$ – aaka Oct 20 '14 at 0:04
2
$\begingroup$

To say that $\lim_{x \to \infty} ( \lfloor x \rfloor -x) = \text{something}$ does not at all imply that $\lim_{x \to \infty} \lfloor x \rfloor = x+\text{something}$. Any expression of the form $\left(\lim\limits_{w\to\text{something}} \text{something} \right)$, if it can be evaluated at all, must come to something not depending on the variable $w$ that is approaching something. That variable is a bound variable.

The function $x\mapsto \lfloor x\rfloor - x$ is periodic with period $1$, i.e. every time $x$ increases by $1$ it starts over and repeats. Such a function cannot have a limit at $\infty$ unless it is a constant function.

| cite | improve this answer | |
$\endgroup$
0
$\begingroup$

Hint: it it were true, then as the sequence $$x_n = n + \frac 12 $$is such as $x_n\to\infty$ then $\lfloor{x_n} - x_n\rfloor \to 0$.

But $$ \lfloor{x_n} - x_n\rfloor = n - \left(n+\frac12\right) = -\frac12 $$

| cite | improve this answer | |
$\endgroup$
0
$\begingroup$

Since f: x -> $\lfloor x \rfloor - x$ is not continuous, I'm not sure using a sequence as a counter example is the right thing to do. Need some proving at least

I would prove that f is 1-periodic and continuous on each interval of the forme ]k, k+1[ :

First you prove that : $\lfloor x+1 \rfloor = \lfloor x \rfloor +1 $

Then you get : f(x+1 ) = $\lfloor x+1 \rfloor -(x+1) = \lfloor x \rfloor -x = f(x) $

Then for x,y each element of ]k,k+[ : $\lfloor x \rfloor =\lfloor y \rfloor$

|f(x)-f(y)| = |$\lfloor x \rfloor -x -(\lfloor y \rfloor -y)| = |x-y|$ -> f continuous on ]k,k+1[

So from these properties, f cannot have 0 as a limit when x-> ∞

| cite | improve this answer | |
$\endgroup$

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.