A function $f:A \subseteq \mathbb{R} \rightarrow \mathbb{R}$ is uniformly continuous if $\forall \epsilon>0$ $\exists \delta>0$ such that $\forall x \in A and \forall y \in A$ with $|x-y|\leq \delta$ we have $|f(x)-f(y)|\leq \epsilon$

Is the negation:

There exists and $\epsilon>0$ such that $\forall \delta>0$ there exists $x\in A$ or there exists $y\in A$ such that $|x-y|\leq \delta$ and $|f(x)-f(y)|\geq \epsilon$?

Show that $f:(0,1)\rightarrow \mathbb{R}$ given by $f(x)=\frac{1}{x}$ is not uniformly continuous.

proof: Let $\epsilon=1$ let $\delta>0$ arbitrary. Set $x\in (0,1)$ to be such that $x<\frac{\delta}{1+\delta}$ and $y= x+\delta$ . Then $|\frac{1}{x}-\frac{1}{x+\delta}|$ $=$ $|\frac{\delta}{x(x+\delta)}|$ $\geq$ $\frac{\delta}{x(1+\delta)}>1$

Is the proof correct? The only question I have is why is $y=x+\delta$ guaranteed to be in $(0,1)$?

  • $\begingroup$ Indeed, you are right, it is not correct. $\endgroup$ – Alexey Oct 28 '19 at 22:03
  • $\begingroup$ If you have found a counterexample for $\delta$,then it is also a counterexample for $\delta' > \delta$. So, you can always assume that $\frac{\delta}{1+\delta}+\delta < 1$. $\endgroup$ – amsmath Oct 28 '19 at 22:12
  • $\begingroup$ @amsmath is the negation correct though? $\endgroup$ – monoidaltransform Oct 28 '19 at 22:13
  • $\begingroup$ The "or there exists" does not make sense. Think about it. $\endgroup$ – amsmath Oct 28 '19 at 22:18
  • $\begingroup$ @amsmath ($\forall x \in A and \forall y \in A$ ) What would be the negation of that statement then? $\exists x\in A or y\in A$, no? $\endgroup$ – monoidaltransform Oct 28 '19 at 22:19

You basically have it. You just need to be a little more careful. You are claiming that for $\epsilon=1$, and for every $\delta>0$, there are $x,y\in (0,1)$ such that$|x-y|<\delta$ and $|f(x)-f(y)|>1.$ You may assume without loss of generality that $\delta<1/2$ because it the claim is true for all such $\delta,$ it will be true for any value of $\delta$ larger than $1/2.$

(Remember, all you need to do is find two numbers in $(0,1)$ whose difference is less than $\delta$ in absolute value. The ones that work for $\delta<1/2$ will also work for $\textit{any}\ \delta\ge 1/2.$ Example: suppose we have $\delta=15$ and you can find $x,y$ such that $|x-y|<1/2$ and $|f(x)-f(y)|>1.$ Then, the $x,y$ work for $\textit{both}$ values of $\delta$ simultaneously because if $|x-y|<1/2$ it is also $<15$).

Now, $|f(x)-f(y)|=\left|\frac{x-y}{xy}\right|$ and we want to choose $x$ and $y$ so that $|x-y|<\delta$ but $\left|\frac{x-y}{xy}\right|>1$, so take $x=\delta$ and $y=2\delta.$ Then, $x$ and $y$ are actually in $(0,1)$ and $\left|\frac{x-y}{xy}\right|=\frac{2}{\delta}>1$, and you are done.

It may be easier to do it with sequences: with $\epsilon=1/2,$ take $\delta_n=1/n$ and find sequences $(x_n)$ and $(y_n)$ such that $|x_n-y_n|\to 0$ but $|f(x_n)-f(y_n)|>1/2.$ Choose $x_n=1/n$ and $y_n=1/n+1$ and check that this assignment works.

  • $\begingroup$ May you elaborate on why I might assume $\delta<\frac{1}{2}$, please? The negation of U.C is to show that it holds for all delta, I cant just show it for some delta. $\endgroup$ – monoidaltransform Oct 28 '19 at 22:31
  • $\begingroup$ Yes, I will add to my answer. $\endgroup$ – Matematleta Oct 28 '19 at 22:32

It is almost correct.Nice work.

To overcome the situation whether $x+\delta \in (0,1)$ you can do this.

Assume that it is uniformly continuous.

Let $x \in (0,1)$

Then for $\epsilon=1$ exists $\delta>0$ such that..e.t.c.

So the ''e.t.c'' part of the proof will be true also for every $\delta_0<\min\{\frac{x-1}{2},\frac{x}{2},\delta\}$

So you can work for $\delta_0$'s the same way you workded since. $x+\delta_0 \in (0,1), \forall \delta_0<\min\{\frac{x-1}{2},\frac{x}{2},\delta\}$

You can also use sequences to prove the statement.

Take $x_n=\frac{1}{n+1}$ and $y_n=\frac{1}{n+2}$

Then $x_n-y_n \to 0$ but $|f(y_n)-f(x_n)|=1 \to 1 \neq 0$

So $f$ is not uniformly continuous.

  • $\begingroup$ please re read what I wrote, please and comment. $\endgroup$ – monoidaltransform Oct 28 '19 at 22:10
  • $\begingroup$ I meant the negation part. $\endgroup$ – monoidaltransform Oct 28 '19 at 22:17
  • $\begingroup$ @topologicalmagician yes the negation is correct. $\endgroup$ – Marios Gretsas Oct 28 '19 at 22:17
  • $\begingroup$ thank you so much $\endgroup$ – monoidaltransform Oct 28 '19 at 22:18
  • $\begingroup$ @topologicalmagician in the negation you must say that there exist x AND y...not x OR y $\endgroup$ – Marios Gretsas Oct 28 '19 at 22:20

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.