0
$\begingroup$

Given a sequence of functions $f_n(x)={{(x-1)^n}\over{1+x^n}} \arctan({n^{x-1}}).$
I have studied its pointwise convergence and I found as set of convergence $E=[0,+\infty)$ and as limit function the null function.
But I have a problem with the uniform convergence: how can I calculate $\;\sup|f_n(x)|\;$ in $E?$

$\endgroup$
  • 2
    $\begingroup$ There is no $y$-dependence in the function. Is this intentional or a typo? $\endgroup$ – Klaus Feb 8 at 12:38
3
$\begingroup$

There is no need to compute the maximum (or minimum) of $f_n$; it is enough to have inequalities. We have $$ |f_n(x)|\le\frac\pi2\,\frac{|x-1|^n}{1+x^n}\le\frac\pi2\,\,\Bigl|1-\frac1x\Bigr|^n\quad\forall x\ge0. $$ Let $A>3/2$. If $3/4\le x\le A$, then $$ |f_n(x)|\le\frac\pi2\,\Bigl|1-\frac1A\Bigr|^n, $$ proving that $f_n$ converges uniformly to $0$ on $[3/4,A]$.

Let's treat now the case $0\le x\le3/4$. Then $n^{x-1}\le n^{-1/4}$, $\arctan(n^{x-1})\le n^{-1/4}$ and $|f_n(x)|\le n^{-1/4}$, proving uniform convergence on $[0,3/4]$.

What happens on $[0,\infty)$? The convergence is not uniform. To prove it, we bound from below $f_n(n)$, $n\in\Bbb N$. First of all, $\arctan(n^{n-1})\ge \pi/4$ for all $n\in\Bbb N$. Next, $$ \frac{|n-1|^n}{1+n^n}\ge\frac{|n-1|^n}{2\,n^n}=\frac12\,\Bigl|1-\frac1n\Bigr|^n\to\frac1e>0. $$ This shows that $f_n(n)\ge c$ for some constant $c>0$, proving that $f_n$ does not converge uniformly to $0$ on $[0,\infty)$.

$\endgroup$
  • 1
    $\begingroup$ It seems to me that the proposed argument is not enough. To conclude that $\left|1-\frac1A\right|^n\rightarrow 0$ as $n\rightarrow +\infty$ you must have $\left|1-\frac1A\right|<1$, that holds if and only if $A>\frac 12$. Thus the proposed estimate shows uniform convergence in $\left[\frac 12 +\delta, \frac 12 + \delta'\right]$ for all $\delta>0$ and $\delta'>\delta$. An extra effort is needed to prove uniform convergence on $[0,\delta]$ for every $\delta>0$. $\endgroup$ – AlessioDV Feb 8 at 17:19
  • $\begingroup$ @AlessioDV Oops! You are right. I will modify my argument. $\endgroup$ – Julián Aguirre Feb 8 at 17:26
  • $\begingroup$ Great answer, i was giving a generic response without trying to solve it . There is indeed no need to compute the maximum $\endgroup$ – Milan Feb 8 at 17:27
0
$\begingroup$

In general you can find the supremum of function by finding its maximum.I dont recommend differentiating the whole function.

You should use the fact that $ arctan $ is bounded by $ \pi/2 $. Then find $max$ of $ {{(x-1)^n}\over{1+x^n}} $ ( the maximum will be a function of n).

[ By multiplying $ \pi/2 $ and $max$ you might get something that is bigger than the actual but it doesnt because matter $ \pi/2 $ can get infront of $lim$.]

$\endgroup$

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.