5
$\begingroup$

I have a general idea of how to prove this but I could use some help with the details.

Basically I see that $f(x)$ is the uniform limit of $f_k(x) = \sum_{n=1}^{k} x^n/n^2$ on $[0,1]$.

Each $f_k$ is continuous, so $f$ is as well since uniform convergence preserves continuity.

Is this proof correct/does it seem sufficient?

Thanks ahead of time.

$\endgroup$
  • 2
    $\begingroup$ The convergence is uniform, by the Weierstrass M-test. I don't see that Weierstrass's approximation theorem is useful here. $\endgroup$ – Lord Shark the Unknown Nov 30 '18 at 4:53
5
$\begingroup$

Note that $|x^n/n^2|\leq 1/n^2$ on $[0,1]$ and $\sum_{n=1}^{\infty} 1/n^2 \lt \infty$. Then by Weierstrass M-test,$\sum_{n=1}^{\infty} x^n/n^2$ converges uniformly. Hence $f$ is continuous.

$\endgroup$
1
$\begingroup$

This can be done by a calculus student, actually, and I can even prove the continuity is uniform, provided said calculus student recalls the mean value theorem. No $M$-test needed here!

Let $x_0, x_1 \in [0,1]$, and let $\epsilon > 0$ be given. We'll find $\delta$ so that $|x_0 - x_1| < \delta \implies |f(x_0) - f(x_1)| < \epsilon$.

To see this, introduce the auxiliary functions $g_n(x) = x^n$

We have:

$$|f(x_1) - f(x_0)| = \bigg|\sum_{n=1}^\infty \frac{x_1^n - x_0^n}{n^2} \bigg| = \bigg| \sum_{n=1}^\infty \frac{g_n(x_1) - g_n(x_0)}{n^2}\bigg|$$

Now $g_n$ are all differentiable, so $g_n(x_1) - g_n(x_0) = g_n'(x)(x_1 - x_0)$ by the Mean Value Theorem, for some $x \in (x_0, x_1)$ (assuming without loss of generality $x_1 > x_0$.

Now we have $g_n'(x) = nx^{n-1}$, so shoving this in to the above equation and observing that now all the terms are positive in the sum, so we drop the bars:

$$|f(x_1) - f(x_0)| \leq \bigg[\sum_{n=1}^\infty \frac{x^{n-1}}{n}\bigg] |x_1 -x_0|$$

But now, the first term is a convergent series, since $x < 1$, and $n > 0$, this expression is bounded above by a geometric series. Calling the limit value $S$, taking $\delta = 1/S$ suffices.

$\endgroup$

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.