Esta lección aún no está traducida. Por ahora se muestra en inglés; la versión en español está en camino.
Taylor series build a function from powers of $x$. Fourier series build a periodic function from sines and cosines, the eigenfunctions of the previous lesson. They describe sound, heat and signals, and they solve the heat and wave equations. Computing coefficients uses integration by parts again and again.
🔑 Idea clave
For $f$ with period $2L$, $$f(x) \sim \frac{a_0}{2} + \sum_{n=1}^{\infty}\left(a_n\cos\frac{n\pi x}{L} + b_n\sin\frac{n\pi x}{L}\right),$$ $$a_n = \frac{1}{L}\int_{-L}^{L} f(x)\cos\frac{n\pi x}{L}\,dx,$$ $$b_n = \frac{1}{L}\int_{-L}^{L} f(x)\sin\frac{n\pi x}{L}\,dx.$$ Here $L$ is half the period. If $f$ is odd, every $a_n = 0$; if $f$ is even, every $b_n = 0$. At a jump, the series converges to the average of the two one-sided values.
💡 Por qué funciona
The building blocks are orthogonal on $[-L, L]$. Using $\cos A\cos B$ $= \frac{1}{2}\left(\cos(A - B) + \cos(A + B)\right)$, the integral $\int_{-L}^{L} \cos\frac{m\pi x}{L}\cos\frac{n\pi x}{L}\,dx$ is a sum of cosines over whole periods, so it is 0 when $m \ne n$ and $L$ when $m = n \ge 1$. The same holds for sines, and every sine-cosine product integrates to 0. Now multiply the series by $\cos\frac{m\pi x}{L}$ and integrate: every term dies except $a_m\cdot L$, which gives the formula for $a_m$. For $n = 0$, $\int_{-L}^{L} 1\,dx = 2L$, which is why the constant term is written $\frac{a_0}{2}$: then one formula covers every $a_n$.
✏️ Ejemplo resuelto: A square wave
Find the Fourier series of $f = -1$ on $(-\pi, 0)$ and $f = 1$ on $(0, \pi)$, period $2\pi$.
Find the series of $f(x) = x^{2}$ on $(-\pi, \pi)$.
$f$ is even: $b_n = 0$. $a_0 = \frac{1}{\pi}\int_{-\pi}^{\pi} x^{2}\,dx = \frac{2\pi^{2}}{3}$, so $\frac{a_0}{2} = \frac{\pi^{2}}{3}$.
Integrating by parts twice, $a_n = \frac{2}{\pi}\int_0^{\pi} x^{2}\cos nx\,dx$ $= \frac{4(-1)^{n}}{n^{2}}$.
Respuesta: $x^{2} \sim \frac{\pi^{2}}{3} - 4\cos x + \cos 2x$ $- \frac{4}{9}\cos 3x + \dots$; at $x = \pi$ this gives $\sum \frac{1}{n^{2}} = \frac{\pi^{2}}{6}$
⚠️ Error común
Using the wrong interval length. $L$ is half the period: for period $2\pi$, $L = \pi$, and for period 6, $L = 3$. The factor in front is $\frac{1}{L}$ (or $\frac{2}{L}$ over half the interval for odd or even functions), and the same $L$ appears inside $\frac{n\pi x}{L}$. Using $\frac{1}{2L}$, or $\frac{1}{\pi}$ when the period is not $2\pi$, scales every coefficient wrongly.
Partial sums of the square wave's series. Add terms: the sum hugs the square wave more closely, passes through the midpoint 0 at each jump, and keeps a small overshoot right next to each jump (the Gibbs phenomenon, about 9% of the jump).The square wave's series at $x = \frac{\pi}{2}$, where $\sin nx$ is $1, 0, -1, 0, \dots$: the terms are $\frac{4}{\pi}\left(1 - \frac{1}{3} + \frac{1}{5} - \dots\right)$, and the partial sums settle on $f\left(\frac{\pi}{2}\right) = 1$.
Práctica
Apruebas con 3 respuestas correctas, las 2 últimas seguidas. Cada respuesta correcta da 3 XP.