0
$\begingroup$

I found this example online:

Let $\{f_{n}\}$ be the sequence of functions on $(0,\infty)$ defined by $f_{n}(x)=\frac{nx}{1+n^{2}x^{2}}$ .This function converges pointwise to zero. Indeed, $(1+n^{2}x^{2})\sim n^{2}x^{2}$ as $n$ gets larger. So, $\underset{n\rightarrow\infty}{\lim}f_{n}(x)=\underset{n\rightarrow\infty}{\lim}f_{n}(x)=\frac{nx}{n^{2}x^{2}}.=\frac{1}{x}\underset{n\rightarrow\infty}{\lim}\frac{1}{n}=0.$

But for any $\epsilon<1/2$ , we have

$|f_{n}(\frac{1}{n})-f(\frac{1}{n})|=\frac{1}{2}-0>\epsilon$ .

Hence$ \{f_{n}\}$ is not uniformly convergent.

  1. By the definition of pointwise convergence: $f_n(x)\rightarrow f(x)$ for all $x\in X$. But this does not seem to happen when $x=\frac{1}{n}$. So how can ${f_n}$ converge pointwise to zero for all $x\in (0,\infty)$ if $f_{n}(\frac{1}{n})=\frac{1}{2}$?

  2. Is it correct to think like this: If $ \{f_{n}\}$ was uniformly convergent, then if someone gave me an $\epsilon<1/2$, I should be able to find and $n\geq N$ so that $|f_{n}(x)-f(x)|$ became smaller than $1/2$ for any $x \in (0,\infty)$, but this doesn't hold for any $x$ because we saw that $|f_{n}(\frac{1}{n})-f(\frac{1}{n})|=\frac{1}{2}-0>\epsilon$, so ${f_n(x)}$ cannot be uniformly convergent.

$\endgroup$
  • $\begingroup$ pointwise means you fix an $x$. $\endgroup$ – Yimin Mar 13 '13 at 18:32
  • $\begingroup$ @Yimin What do you man by "fix an x"? What happens if i fix $x=1/n$? $\endgroup$ – john.abraham Mar 13 '13 at 18:33
  • $\begingroup$ $1/n$ is changing... $\endgroup$ – Yimin Mar 13 '13 at 18:33
  • $\begingroup$ @Yimin Ohh, of course. Thanks! :) $\endgroup$ – john.abraham Mar 13 '13 at 18:34
0
$\begingroup$

Pointwise convergence $f_n(x)\underset{n\to\infty}{\to}{f(x)}$ is convergence by $n$ at every (but fixed) argument $x$.

$\endgroup$

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.