Pointwise convergence of Fourier series

-

Our main question has been: What type of functions can be represented as a Fourier series? We have seen two types of convergence of \(S[f]\) to \(f\). 

  1. Uniform convergence: If \(f\in C_{2\pi}^1\) (continuously differentiable), then \(S[f]\) converges uniformly to \(f\).
  2. \(L^2\) convergence: If \(f\in\mathcal{R}_{2\pi}^{2}\) (square-integrable), then \(S[f]\) converges to \(f\) in the \(L^2\) norm.
These are ``global'' convergence properties, so to speak, in the sense that they characterize the function \(f\) for the whole period \([-\pi, \pi]\).

Now, we consider ``local'' convergence properties such as the problem of pointwise convergence. 


To begin with, we have the following result:

Theorem (du Bois-Reymond)

Let \(f\) be a periodic continuous function. Then, the Fourier series of \(f\) may not converge pointwise.

But we don't really need the continuity of \(f\) for its Fourier series to converge to \(f(x_0)\) at \(x_0\). Thus, let's define the following.

Definition (piecewise continuous and smooth functions)

A function \(f\) with a period \(2\pi\) is said to be piecewise continuous and smooth if there exist finitely many points \(\gamma_1 < \gamma_2 < \cdots < \gamma_l\) in the interval \([-\pi, \pi)\) such that \(f\) and \(f'\) are continuous and bounded in each interval \((\gamma_{j}, \gamma_{j+1})\) and in \((\gamma_l - 2\pi, \gamma_1)\).

If \(f\) is piecewise continuous and smooth, the left limit \(f(\gamma_j-0) = \lim_{x\to \gamma_j - 0}f(x)\) and the right limit \(f(\gamma_j + 0) = \lim_{x\to \gamma_j + 0}f(x)\) exist at each \(\gamma_j\). In fact, if \(x_1, x_2 \in (\gamma_{j-1}, \gamma_{j})\), then
\[|f(x_1) - f(x_2)| = \left|\int_{x_1}^{x_2}f'(\xi)d\xi\right| \leq M_{j-1}|x_1 - x_2|\]
where \(M_{j-1} = \sup_{\xi\in(\gamma_{j-1},\gamma_{j})}|f'(\xi)|\). Thus,
\[\lim_{x_1,x_2\to \gamma_j - 0}|f(x_1) - f(x_2)| = 0.\]
Similarly, \(\lim_{x_1,x_2\to \gamma_{j-1} + 0}|f(x_1) - f(x_2)| = 0\).

In particular, \(f\) is uniformly continuous on each subinterval \((\gamma_{j-1}, \gamma_j)\).

See also: Uniform continuity (Wikipedia)

We use the following lemma without proof.

Lemma (Riemann-Lebesgue Lemma)

Let \(g(x)\) be a function on an arbitrary interval \(I = (a, b)\) that is continuous except for finitely many points and integrable on \(I\), that is,
\[\int_a^b|g(x)|\,dx < +\infty.\]
Then,
\[\int_a^be^{i\lambda x}g(x)dx \to 0 ~~~ (\lambda \to \pm\infty).\]
Remark. If we adopt the Lebesgue integral, the required conditions are that \(g\) is measurable and integrable on \(I\). □
Proof. Omitted. ■
See also: Riemann-Lebesgue Lemma (Wikipedia)

Theorem (Pointwise convergence of Fourier series)

If the function \(f\) is piecewise continuous and smooth, then, for each \(x_0\), the Fourier series of \(f\) converges to
\[\frac{1}{2}\{f(x_0 - 0) + f(x_0 + 0)\}.\]
Proof. Let \(S_n\) be the partial sum of the Fourier series of \(f\). Using the Dirichlet kernel (which is an even function),
\[\begin{eqnarray*} S_n(x_0) &=& \int_{x_0-\pi}^{x_0+\pi}D_n(y - x_0)f(y)dy\\ &=& \int_{-\pi}^{\pi}D_n(y)f(x_0 + y)dy\\ &=& \int_{-\pi}^{0}D_n(y)f(x_0 + y)dy + \int_0^{\pi}D_n(y)f(x_0 + y)dy\\ &=& \int_{0}^{\pi}D_n(y)\{f(x_0 + y) + f(x_0 - y)\}dy. \end{eqnarray*}\]
Since the Dirichlet kernel is an even function and normalized,
\[2\int_{0}^{\pi}D_n(y)dy = \int_{-\pi}^{\pi}D_n(y)dy = 1.\]
Therefore, it suffices to show (why?) that
\[\begin{eqnarray*} J_1 &\equiv& \int_0^{\pi}D_n(y)\{f(x_0 + y) - f(x_0 + 0)\}dy \to 0,\\ J_2 &\equiv& \int_0^{\pi}D_n(y)\{f(x_0 - y) - f(x_0 - 0)\}dy \to 0. \end{eqnarray*}\]
Since \(f\) is piecewise continuous and smooth (by assumption), there exists a \(\delta > 0\) such that \(f'(x)\) is continuous and bounded for all \(x \in (x_0, x_0 + \delta)\). Let us define
\[\begin{eqnarray*} M_{\delta} &=& \sup_{x \in (x_0, x_0 + \delta)}|f'(x)|,\\ g(y) &=& f(x_0 + y) - f(x_0 + 0). \end{eqnarray*}\]
Then, for \(0 < y < \delta\),
\[|g(y)| = \left|\int_{0}^{y}f'(x_0 + t)dt\right| \leq \int_{0}^{y}|f'(x_0 + t)|dt \leq M_{\delta}y.\]
Therefore,
\[\int_{0}^{\delta}\left|\frac{1}{\sin\frac{y}{2}}g(y)\right|dy \leq M_{\delta}\int_0^{\delta}\frac{y}{\sin\frac{y}{2}}dy < +\infty.\]
For \(\delta < y < \pi\),
\[\sin\frac{y}{2} \geq \sin\frac{\delta}{2} > 0\]
so that
\[\int_{\delta}^{\pi}\left|\frac{1}{\sin\frac{y}{2}}g(y)\right|dy \leq \frac{1}{\sin\frac{\delta}{2}}\int_{\delta}^{\pi}|g(y)|dy < +\infty.\]
Therefore, \(g(y)/\sin(y/2)\) is integrable on \((0, \pi)\). By the Riemann-Lebesgue lemma, 
\[J_1 = \frac{1}{2\pi}\int_0^{\pi}\left(\sin\left(n + \frac{1}{2}\right)y\right)\cdot\frac{g(y)}{\sin\frac{y}{2}}dy \to 0.\]
We can prove that \(J_2 \to 0\) in a similar manner (exercise!). ■

Remark. If \(f\) is continuous at \(x_0\), its Fourier series at \(x_0\) converges to \(f(x_0)\). If we define the function \(\tilde{f}\) by
\[\tilde{f}(x) = \begin{cases} f(x) & \text{if $f$ is continuous at $x$},\\ \frac{1}{2}\{f(x-0) + f(x+0)\} & \text{if $f$ is discontinuous at $x$}, \end{cases}\]
then, the Fourier series of \(\tilde{f}\) converges to \(\tilde{f}\) at each point. □

Remark. The above theorem assumes that \(f'\) is piecewise continuous and bounded. This condition can be relaxed in various ways. For example, we may assume instead that
  • \(f'\) is piecewise continuous and square-integrable.
□

Comments

Post a Comment

Popular posts from this blog

Open sets and closed sets in \(\mathbb{R}^n\)

-
Image
Open sets In \(\mathbb{R}\), we have the notion of an open interval such as \((a, b) = \{x \in \mathbb{R} | a < x < b\}\). We want to extend this idea to apply to \(\mathbb{R}^n\). We also introduce the notions of bounded sets and closed sets in \(\mathbb{R}^n\). Recall that the \(\varepsilon\)-neighbor of a point \(x\in\mathbb{R}^n\) is defined as \(N_{\varepsilon}(x) = \{y \in \mathbb{R}^n | d(x, y) < \varepsilon \}\) where \(d(x,y)\) is the distance between \(x\) and \(y\). Definition (Open set) A subset \(U\) of \(\mathbb{R}^n\) is said to be an  open set  if the following holds: \[\forall x \in U ~ \exists \delta > 0 ~ (N_{\delta}(x) \subset U).\tag{Eq:OpenSet}\] That is, for every point in an open set \(U\), we can always find an open ball centered at that point, that is included in \(U\). See the following figure. Perhaps, it is instructive to see what is  not  an open set. Negating (Eq:OpenSet), we have \[\exists x \in U ~ \forall \delta > 0 ~ (N_{\delta}(x) \not
Read more

Euclidean spaces

-
Image
We would like to study multivariate functions (i.e., functions of many variables), continuous multivariate functions in particular. To define continuity, we need a measure of "closeness" between points. One measure of closeness is the Euclidean distance. The set \(\mathbb{R}^n\) (with \(n \in \mathbb{N}\)) with the Euclidean distance function is called a Euclidean space. This is the space where our functions of interest live. The real line is a geometric representation of \(\mathbb{R}\), the set of all real numbers. That is, each \(a \in \mathbb{R}\) is represented as the point \(a\) on the real line. The coordinate plane , or the \(x\)-\(y\) plane , is a geometric representation of \(\mathbb{R}^2\), the set of all pairs of real numbers. Each pair of real numbers \((a, b)\) is visualized as the point \((a, b)\) in the plane. Remark . Recall that \(\mathbb{R}^2 = \mathbb{R}\times\mathbb{R} = \{(x, y) | x, y \in \mathbb{R}\}\) is the Cartesian product of \(\mathbb{R}\) with i
Read more

Newton's method

-
Image
Newton's method (the Newton-Raphson method) is a very powerful numerical method for solving nonlinear equations.  Suppose we'd like to solve a non-linear equation \(f(x) = 0\), where \(f(x)\) is a (twice) differentiable non-linear function. Newton's method generates a sequence of numbers \(c_1, c_2, c_3, \cdots\) that converges to a solution of the equation. That is, if \(\alpha\) is a solution (i.e., \(f(\alpha) = 0\)) then,  \[\lim_{n\to\infty}c_n = \alpha,\] and this sequence \(\{c_n\}\) is generated by a series of linear approximations of the function \(f(x)\). Theorem (Newton's method) Let \(f(x)\) be a function that is twice differentiable on an open interval \(I\) that contains the closed interval \([a, b]\) (i.e., \([a,b]\subset I\)) and satisfy the following conditions: \(f(a) < 0\) and \(f(b) > 0\); For all \(x\in [a, b]\), \(f'(x) > 0\) and \(f''(x) > 0\). Let us define the sequence \(\{c_n\}\) by \[ \begin{eqnarray} c_1 &
Read more