I've been trying to get my head around this problem for quite some time by now. I want to prove that $$f(x) := \left|\frac{x-1}{x^2+1}\right|$$ is continuous for $$x_0 = -1$$

Now, in order to prove this, I want to use to ε-δ defintion which states that for any given $\varepsilon$, there exists $\delta>0$ such that

$$\left|x-x_0\right| < \delta \Rightarrow \left|f(x)-f(x_0)\right|<\varepsilon$$

I'm fairly new to the concepts of analysis; but I understand that in order to conduct these sorts of proofs, you first need to spend a few minutes to find out a reasonable value for $\delta$. Unfortunately, I can't find any particular "rules" on how to do that. So, I first tried to fill in some gaps for my particular problem:

Given $\varepsilon > 0$, we can find $\delta > 0$ such that, if $\left|x-(-1)\right| = \left|x+1\right| < \delta$, then $$\left|\frac{x-1}{x^2+1}-(-1)\right|=\left|\frac{x-1}{x^2+1}+1\right|=\left|\frac{x²+x}{x^2+1}\right|<\varepsilon$$

My plan is now to take care of the implication first. That is, I want to paraphrase $\left|\frac{x²+x}{x^2+1}\right|<\varepsilon$ in a way that makes sure that the actual $\varepsilon > 0$ chosen does not matter at all. So, in the end, I think manipulating $\left|\frac{x²+x}{x^2+1}\right|<\varepsilon$ to look like $\left|x+1\right| < ε \cdot \varphi$ would be my goal, right? Because then, if we let $\delta := \varepsilon \cdot \varphi$, the implication holds.

I hope everything to this point was correct. Now, I find

$$\left|\frac{x²+x}{x^2+1}\right| = \left|\frac{x(x^2+1)}{x+1}\right| = \left|x+1\right|\left|\frac{x}{x^2+1}\right|<\varepsilon$$ which is why I would want to let $\delta := \frac{\varepsilon}{\left|\frac{x}{x^2+1}\right|}$ (which looks quite ugly, point taken).

Is my choice of $\delta$ okay? I never saw a $\delta$ contain $x$ as a variable before, my professor always somehow managed to set $\delta$ to $\varepsilon$ times some constant $\varphi$, but I can't seem to do it.

Can anyone check my current steps and help me find a constant $\varphi$, if this is necessary for the proof I'm building up?

  • 2
    $\begingroup$ Delta cannot involve $x$. Remember that $\delta$ defines the range of $x$-values being considered, so it doesn't make sense for it to have $x$ in it. $\endgroup$ – user142299 May 12 '14 at 23:58

Note that

$$\left|\frac{x}{x^2+1}\right|<1, \forall x \in \mathbb{R}.$$ So, taking $\epsilon=\delta$ you have

$$|x+1|<\delta \Rightarrow \left|\frac{x^2+x}{x^2+1}\right| = \left|\frac{x(x+1)}{x^2+1}\right| = \left|x+1\right|\left|\frac{x}{x^2+1}\right|<|x+1|<\delta=\epsilon.$$

| cite | improve this answer | |
  • $\begingroup$ Wonderful, I didn't even see that :) Small correction: It should be $\left|\frac{x(x+1)}{x+1}\right|$ in the middle. $\endgroup$ – chiru May 13 '14 at 0:09
  • 1
    $\begingroup$ This solution is clever, but the one by @Omnomnomnom is much better for understanding how to do this type of problem in general. Setting delta equal to the minimum of a constant and some function of epsilon (in this case, $\delta = min\{1, \epsilon/2\}$) is a standard approach for these types of problems $\endgroup$ – ahuff44 May 13 '14 at 0:12

You don't need the best $\delta$, you just need a $\delta$ that works.

In particular, when $|x+1| < \delta$, we need to have $$ \left|x+1\right|\left|\frac{x}{x^2+1}\right|<\varepsilon $$ Or in other words, we need to have $$ \delta \left|\frac{x}{x^2+1}\right|<\varepsilon $$ But let's presuppose that we're going to set $\delta < 1$. If we already know that $|x + 1| < 1$, then we know $-2 < x < 0$. So, we can say that $$ \left|\frac{x}{x^2+1}\right| = \frac{|x|}{|x^2 + 1|} < \frac{2}{x^2+1} < \frac{2}{0^2+1} = 2 $$ So, presupposing $\delta < 1$, it's enough to have $$ \delta \cdot 2 < \varepsilon $$ So, one satisfactory choice of $\delta$ will be $\delta = \min\{1,\epsilon/2\}$.

| cite | improve this answer | |

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.