Term by Term Differentiability in the context of Uniform Convergence

817 Views Asked by At

I'm not sure how differentiability works with uniform convergence. My book says that we can show this (calculation wise) $$\varepsilon (x,a) = \sum_{k=1}^{\infty} E_{k}(x,a)$$

for some $x$ and $a$.

In this context, I'm not sure how to go about doing this:

Prove that $$f(x)=\sum_{n=1}^{\infty}\frac{1}{n^{2}+x^{2}}$$ is differentiable for all values of $x$.

If someone could give me a hand, that would be great. My book has no examples in this topic, so I just chose one of the first questions from it.

(If more context is needed, please let me know. This is my first stackexchange question.)

1

There are 1 best solutions below

3
On BEST ANSWER

You cannot always differentiate term-by-term but in this case we can. Let $f_n(x)=1/(n^2+x^2)$, let $F_k(x)=\sum_{n=1}^kf_n(x)$ , let $$f(x)=\sum_{n=1}^{\infty}f_n(x)=\lim_{k\to \infty}F_k(x).$$ Let $g_k(x)=\sum_{n=1}^kf'_n(x)$ and let $$g(x)=\sum_{n=1}^{\infty}f'_n(x)=\lim_{k\to \infty}g_k(x).$$ Now each $g_k$ is continuous. If you can show that $g_k$ converges uniformly to $g$ on any bounded real set as $k\to \infty$, then you know $g$ is continuous and that $$f(x)-f(a)=\lim_{k\to \infty}(F_k(x)-F_k(a))=\lim_{k\to \infty}\int_a^xg_k(y) dy=\int_a^xg(y)dy$$ for any $x,a.$ Now since $g$ is continuous we have (with any fixed $a\ne x$),$$f'(x)=g(x).$$