1

This is problem 3.2.7. from Problems in Real Analysis, Radulescu et. al., I'm stuck on the very last part.

The problem is defined such that $x$ is in the open interval $(0,1)$. From $\lim\limits_{x\to 0} \frac{f(x)-f\left(x/2\right)}x=0$, given an $\epsilon$, we have: $$|f(x)-f\left(\frac{x}{2}\right)|<\varepsilon x$$ from which we can obtain $|f(\frac{x}{2^n})-f(\frac{x}{2^{n+1}})|< \varepsilon\frac{x}{2^n}$ and using the triangle inequality we get $$|f(x)-f\left(\frac{x}{2^n}\right)| \le 2\varepsilon x.$$ If we take $n\to\infty$, since $\lim\limits_{x\to 0} f(x)=0$, we get: $$|f(x)|\le 2\varepsilon |x|$$ $$\frac{|f(x)|}{|x|} \le 2\varepsilon$$

My question is, how come this implies that $\lim \limits_{x \to 0} \frac{f(x)}{x} = 0$? Don't we still need to go from $\frac{|f(x)|}{|x|} \leq 2\varepsilon$ to $\frac{|f(x)|}{|x|} < \varepsilon$?

  • see accepted answer on https://math.stackexchange.com/questions/89575/if-fx-and-f2x-fx-x-have-limit-0-as-x-to-0-then-fx-x-to-0/89620#89620 for a neater proof, if it is of interest to anyone – shintuku Jul 28 '21 at 13:04
  • 2
    It doesn't matter. Epsilon is a quantity that you can take as small as you need, and so is $2\epsilon$. A while back, I gave a proof on this matter in an answer. Hold on – Stefan Octavian Jul 28 '21 at 13:05
  • @StefanOctavian I was expecting as much, I would appreciate that link very much! – shintuku Jul 28 '21 at 13:06
  • See this question https://math.stackexchange.com/questions/3983841/thoughts-about-the-use-of-epsilon-delta-language-in-proofs-involving-limit/3983858#3983858 – Stefan Octavian Jul 28 '21 at 13:07
  • @StefanOctavian thank you very much! – shintuku Jul 28 '21 at 13:08
  • 1
    See https://math.stackexchange.com/a/1864092/72031 for a more general version. – Paramanand Singh Jul 30 '21 at 07:34

1 Answers1

1

Note that $\frac{|a|}{|b|} = \left|\frac{a}{b}\right|,$ so the inequality you have and the inequality you want are equivalent.