0
$\begingroup$

I got this power series $$\sum_{n=1}^{\infty} x^2 e^{-nx} $$ And i need to prove uniform convergence when $x \in[0;1]$ and find this sum. I have proven uniform convergence, but i have no idea how to find this sum. It's clear, that $\sum_{n=1}^{\infty} x^2 e^{-nx}$ is a geometric series $\frac{x^2 e^{-x}}{1-e^{-x}}$. Is it correct that $\sum_{n=1}^{\infty} x^2 e^{-nx}$ equals $\int_0^1 \frac{x^2 e^{-x}}{1-e^{-x}}$?

$\endgroup$
  • 2
    $\begingroup$ Do you need the value of the series or the value of an integral over this series? You have not clearly stated why you would need the $\int_0^1$ part in your question. $\endgroup$ – Formyer Oct 9 '17 at 2:05
  • $\begingroup$ @Formyer i need find value of the series $\endgroup$ – Shmuser Oct 9 '17 at 2:06
  • 1
    $\begingroup$ Seems to me you already have all the answers (convergence, geometric serie, the value of the sum).What is it, you need exactly? $\endgroup$ – zwim Oct 9 '17 at 2:08
  • $\begingroup$ @Formyer I thought that value of the series and the value of an integral over this series are the same, aren't they? $\endgroup$ – Shmuser Oct 9 '17 at 2:09
  • 1
    $\begingroup$ @user401689 No. The series depends on $x$ but the integral is a specific number, so even without computing either of them, you can see they can't be the same. I'm not sure why you even integrated it, in the first place. $\endgroup$ – Jam Oct 9 '17 at 2:14
3
$\begingroup$

You already computed the correct answer

$$\sum_{n=1}^\infty x^2e^{-nx}=x^2\frac{e^{-x}}{1-e^{-x}}.$$

You seem to conflate a few different concepts here: For one,

$$\sum_{n=1}^\infty x^2e^{-nx}=\lim_{k\rightarrow\infty}\sum_{n=1}^k x^2e^{-nx}$$

(a series is the limit of its partial sums) is the definition of a series.

Then, you have some "similarities" between series and integrals that do not factor into this problem: For starters, the Riemann integral approach:

$$\sum_{i=1}^N f(x_i)\Delta x_i\rightarrow\int_a^bf(x)\text dx$$

for a partition $a=x_1<...<x_N=b$ of an interval $[a,b]$, with $\Delta x_i=x_{i+1}-x_i$. The limit converges for any sequence of partitions as long as the mesh size $\max_{i=1,...,N-1}(x_{i+1}-x_i)$ converges towards zero.

Then, there is the integrability criterion / integral test that lets you determine whether a series converges at all by comparing it to the integral over its summation sequence. That only holds for non-negative monotonely decreasing functions, though, and is a comparison, not an equality.

Again, none of this has anything to do with the answer to your question, but might clear up some concepts you seem to have difficulty grasping.

$\endgroup$

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.