With $\displaystyle c_n : t \mapsto \sqrt{\frac{2}{\pi}}\cos\left(nt\right)$ i've shown that $\underset{n \in \mathbb{N}}{\text{Span}}\left(c_n\right)$ was dense in the set $G$ of real continuous functions on $\left[0,\pi\right]$. ( not necessarily $2 \pi$ periodic ). And i've a theorem that says that for $x \in G$ the orthogonal projection $p_{F_n}$ on $F_n=\text{Span}\left(c_0,c_1,\dots,c_n\right)$ satisfies $$ p_{F_n}\left(x\right) \underset{n \rightarrow +\infty}{\rightarrow}x $$ Am i wrong ? How can I use this to show that all function $2\pi$-periodic even continuous function on $\left[0,\pi\right]$ satisfies $$ f=\sum_{n=0}^{+\infty}\langle f,c_n\rangle c_n $$ with $\displaystyle \langle f,g \rangle =\int_{0}^{\pi}f(t)g(t)\text{d}t$


Your last point $f=\sum_{n=0}^{+\infty}\langle f,c_n\rangle c_n$ for even continuous functions is surely not true at it exist continuous real maps with diverging Fourier series.

  • $\begingroup$ If we have also $f$ is $2\pi$ périodic then it is true no ? $\endgroup$ – Atmos May 20 '18 at 13:39
  • $\begingroup$ No . For the same reason. Not all continuous functions are developpable in Fourier series. $\endgroup$ – mathcounterexamples.net May 20 '18 at 16:31
  • $\begingroup$ Can i have an example of continuous, $2\pi$-periodic and even function that is not developpable in Fourier series ? Seeing your pseudo, it is easy for you $\endgroup$ – Atmos May 20 '18 at 19:09
  • $\begingroup$ @Atmos : You could qualify by saying that the series converges in the $L^2$ norm. $\endgroup$ – DisintegratingByParts May 21 '18 at 22:54

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.