I want to know whether this function series in uniformly converges or not. I can recognize Leibniz here, so if I'll be able to prove that $\lim_{n \to \infty} f_n(x)=f(x)=0$ the function series will be uniformly converged to 0.

How should I do that?


  • $\begingroup$ One should have $g(x) \sum_{n=0}^\infty h(x)^n = g(x)/(1-h(x))$ for all $x$ such that $|h(x)| < 1$. In this case $g(x) = -x^2$, and $h(x)=-1/(1+x^2)$. So you sure it converges to $0$ for all $x$? $\endgroup$ – p.s. Dec 7 '11 at 10:05
  • $\begingroup$ @p.s. I´m sorry but I don´t think I have got your comment. May I ask you to be a little more explicit? In specific I don´t understand why you say: "One should have... ...for all $x$ such that $|h(x)|\leq 1$". $\endgroup$ – Giovanni De Gaetano Dec 7 '11 at 10:16
  • $\begingroup$ This is an answer. I write it down first as a comment because I don´t feel confident in its exactness. As Jozef pointed out it is enough to prove the limit: $\lim_{n\rightarrow\infty} f_n(x)=0$ for each $x$, in order to have that the series is uniformly convergent (not to zero!). Then a computation shows that $max(f_n(x))= \frac{2}{n-2}(1+\frac{2}{n-2})^{-n} \leq \frac{1}{n}\cdot k \rightarrow 0$. Now, by Leibniz Criterion (it wouldn´t be true otherwise) the original series is uniformly convergent. Please give me a feedback, so (if correct) I can post it as an answer. $\endgroup$ – Giovanni De Gaetano Dec 7 '11 at 10:44
  • $\begingroup$ @Student73 I'm sorry, I didn't see your comment. I don't think showing pointwise convergence is sufficient, though (?). I deleted my answer. If you like, you can post yours (your inequality shows the convergence is uniform) $\endgroup$ – David Mitra Dec 7 '11 at 11:44
  • $\begingroup$ @David Mitra. I´m really sorry you deleted your answer! I didn´t mean to be ironic or sarcastic! Actually I think it is correct that you answer the question, for two reasons. 1) Who really answer the question is who knows that the answer is correct, I guessed but I didn´t know! 2) I´m always a little confused about different kinds of convergence and I don´t want take the risk to insert an error in the answer. Anyway I want to really thank you for your correctness and politeness! :) $\endgroup$ – Giovanni De Gaetano Dec 7 '11 at 12:23

Your series is a geometric series; therefore the partial sums $s_n:=\sum_{k=0}^{n-1}\ldots$ can be computed explicitly. One obtains

$$s_n={x^2(1+x^2) \over 2+x^2}\ \left(1-\Bigl({-1\over 1+x^2}\Bigr)^n\right)\ ;$$

so it is enough to prove that $g_n(x):={x^2\over (1+x^2)^n}$ converges to $0$ uniformly with $n\to\infty$. To this end put $x^2=:u$ and investigate

$$h_n(u):=u(1+u)^{-n}\qquad (0\leq u<\infty)$$

instead. It is easy to verify that for $n\geq2$ the function $h_n$ assumes its maximum at $u_n={1\over n-1}$; and the value there is

$$h(u_n)={1\over n-1}\ \Bigl(1+{1\over n-1}\Bigr)^{-n}\ .$$

As the second factor on the right converges to ${1\over e}$ it follows that $h_n(u)\leq {C\over n}$ for some $C$ and all $u\geq 0$, $n\geq2$.

| 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.