A proof about approximations in L1

51 Views Asked by At

I'm getting a little confused about the next exercise:

Suppose that $f_{k}\in L^1[-\pi,\pi]$ and $f\in L^1[-\pi,\pi]$ are such that $f_k \to f$ when $k\to \infty $.

Prove that $\hat{f_k} \to \hat{f}$ uniformly, where $\hat{f_k}$ and $\hat{f}$ are the Fourier coefficents of $f_k$ and $f$ respectively.

Mi attempt was:

For an $\varepsilon>0$ we know that $\exists N \in \mathbb{N}$ such that $|f_k(\theta)-f(\theta)|<\varepsilon$ for every $k>N$ and $\theta\in[-\pi,\pi]$. Then, notice the following:

$\hat{f_k}-\hat{f}=\frac{1}{2\pi}\int_{-\pi}^{\pi} f_k(\theta) e^{-in\theta}dx-\frac{1}{2\pi}\int_{-\pi}^{\pi} f(\theta) e^{-in\theta}dx=\frac{1}{2\pi}\int_{-\pi}^{\pi} [f_k(\theta)-f(\theta)] e^{-in\theta}dx\frac{1}{2\pi}\leq\frac{1}{2\pi}|\int_{-\pi}^{\pi} [f_k(\theta)-f(\theta)] e^{-in\theta}dx|\leq\frac{1}{2\pi}\int_{-\pi}^{\pi} |f_k(\theta)-f(\theta)| e^{-in\theta}dx\leq\frac{1}{2\pi}\int_{-\pi}^{\pi} \varepsilon e^{-in\theta}dx=\frac{\varepsilon}{2\pi}\int_{-\pi}^{\pi} e^{-in\theta}dx$

My problem is that the final integral is equal to $\frac{2sin(n\pi)}{n}=0$ since $n\in\mathbb{N}$, and I don't like that. Do you see any mistake? or Can you see another way to prove this?

Any help is appreciated, thank you.

1

There are 1 best solutions below

0
On BEST ANSWER

Your inequalities don't make sense. The argument is quite simple: $|\int f_n(t)e^{-int} dt-\int f(t)e^{-int}dt|\leq \int |f_k(t)-f (t)||e^{-int}|dt \to 0$ since $|e^{-int}|=1$ for all $t$ and $n$.

[You did not in what sense $f_k \to f$. I am assuming that $\int |f_n-f| \to 0$].