Vector space of functions

-
We study the collection of functions \(\mathcal{R}_{2\pi}^2\) (square-integrable functions with period \(2\pi\)) as a vector space. We define the \(L^2\) norm and \(L^2\) inner product on this vector space so that we can investigate the ``geometric'' structure of the space of functions. 


Let us show that \(\mathcal{R}_{2\pi}^{2}\), the set of square-integrable functions with period \(2\pi\), is a vector space over \(\mathbb{C}\). First, we need to define addition and scalar multiplication. Let \(f, g\in \mathcal{R}_{2\pi}^2\). We define \(f + g \in \mathcal{R}_{2\pi}^2\) by

\[(f+g)(x) = f(x) + g(x), ~ x \in \mathbb{R}.\tag{eq:add}\]
Note that the ``\(+\)'' on the left-hand side is defined between the two functions \(f\) and \(g\), whereas the ``\(+\)'' on the right-hand side is the addition between two complex numbers \(f(x)\) and \(g(x)\). Next, we define scalar multiplication. Let \(\alpha\in \mathbb{C}\) and \(f \in \mathcal{R}_{2\pi}^2\). We define \(\alpha f\) by
\[(\alpha f)(x) = \alpha f(x).\tag{eq:scale}\]
Note that the product on the left-hand side is between the scalar \(\alpha\) and the function \(f\), whereas the product on the right-hand side is between the two complex numbers \(\alpha\) and \(f(x)\).

Lemma 

\(\mathcal{R}_{2\pi}^2\) is a vector space with vector addition (eq:add) and scalar multiplication (eq:scale).
Proof. We show that \(\mathcal{R}_{2\pi}^2\) satisfies all the axioms of the vector space. In the following, \(f, g, h\in\mathcal{R}_{2\pi}^2\) and \(\alpha, \beta \in \mathbb{C}\).

1. \(\mathcal{R}_{2\pi}^2\) is closed under vector addition.

Suppose \(f, g \in \mathcal{R}_{2\pi}^2\). We show \(f + g\in\mathcal{R}_{2\pi}^2\).

By assumption, \(\int_{\pi}^{\pi}|f(x)|^2\,dx < +\infty\) and \(\int_{-\pi}^{\pi}|g(x)|^2\,dx < +\infty\). Thus,

\[\begin{eqnarray} \int_{-\pi}^{\pi}|f(x)+g(x)|^2\,dx &\leq&\int_{-\pi}^{\pi}(|f(x)| + |g(x)|)^2\,dx ~ (\text{the triangle inequality})\\ & = & \int_{-\pi}^{\pi}|f(x)|^2\,dx + \int_{-\pi}^{\pi}|g(x)|^2\,dx + 2\int_{-\pi}^{\pi}|f(x)||g(x)|\,dx\\ &\leq&\int_{-\pi}^{\pi}|f(x)|^2\,dx + \int_{-\pi}^{\pi}|g(x)|^2\,dx + 2\int_{-\pi}^{\pi}\frac{|f(x)|^2 + |g(x)|^2}{2}\,dx\\ &=& 2\left(\int_{-\pi}^{\pi}|f(x)|^2\,dx + \int_{-\pi}^{\pi}|g(x)|^2\,dx \right)\\ & < & +\infty ~ (\text{by assumption}). \end{eqnarray}\]

2. Addition is commutative.

For any \(f, g\in \mathcal{R}_{2\pi}^2\),

\[(f+g)(x) = f(x) + g(x) = g(x) + f(x) = (g+f)(x).\]

3. Addition is associative.

For any \(f, g, h \in \mathcal{R}_{2\pi}^2\),
\[\begin{eqnarray}((f+g)+h))(x) &=& (f+g)(x) + h(x)\\ &=& (f(x) + g(x)) + h(x)\\ &=& f(x) + (g(x)+h(x))\\ &=& f(x) + (g+h)(x)\\ & =& (f+(g+h))(x). \end{eqnarray}\]

4. The existence of the zero element (additive identity).

Let \(\mathbf{0}(x) = 0\) be the identically zero function. Clearly, \(\mathbf{0}\in \mathcal{R}_{2\pi}^2\), and for any \(f\in\mathcal{R}_{2\pi}^2\),
\[(\mathbf{0} + f)(x) = \mathbf{0}(x) + f(x) = f(x).\]

5. Additive inverses exist.

For any \(f\in\mathcal{R}_{2\pi}^{2}\), define \(-f\) by \((-f)(x) = -f(x)\). Clearly, \(-f\in\mathcal{R}_{2\pi}^2\), and 
\[(f + (-f))(x) = f(x) + (-f(x)) = 0 = \mathbf{0}(x).\]

6. \(\mathcal{R}_{2\pi}^2\) is closed under scalar multiplication.

For each \(\alpha \in\mathbb{C}\) and \(f \in\mathcal{R}_{2\pi}^2\),
\[\int_{-\pi}^{\pi}|\alpha f(x)|^2\,dx=|\alpha|^2\|f\|^2 < +\infty.\]

7. Multiplication by the sum of scalars: \((c+d)f = cf + df\).

For each \(f\in \mathcal{R}_{2\pi}^2\) and \(\alpha, \beta \in \mathbb{C}\),
\[((\alpha+\beta)f)(x) = (\alpha+\beta)f(x) = \alpha f(x) + \beta f(x) = (\alpha f)(x) + (\beta f)(x) = (\alpha f + \beta f)(x).\]
Thus, \((\alpha +\beta )f = \alpha f + \beta f\).

8. Scalar multiplication of the sum of vectors: \(\alpha (f + g) = \alpha f + \alpha g\).

For all \(f, g\in\mathcal{R}_{2\pi}^2\), and \(\alpha \in\mathbb{C}\),
\[(\alpha (f + g))(x) = \alpha (f(x) + g(x)) = \alpha f(x) + \alpha g(x) = (\alpha f+\alpha g)(x).\]

9. The existence of the multiplicative identity scalar.

For any \(f\in\mathcal{R}_{2\pi}^{2}\), \(1 \in \mathbb{C}\),
\[(1\cdot f)(x) = 1\cdot f(x) = f(x).\]

10. Scalar multiplication is associative: \((\alpha\beta )f = \alpha(\beta f)\).

For any \(\alpha, \beta\in \mathbb{C}\), and \(f \in \mathcal{R}_{2\pi}^2\), 
\[((\alpha\beta)f)(x) = (\alpha\beta)f(x) = \alpha(\beta f(x)) = \alpha(\beta f)(x) = (\alpha (\beta f))(x).\]

Thus, \(\mathcal{R}_{2\pi}^{2}\) is a vector space over \(\mathbb{C}\). ■

Definition (\(L^2\)-norm, mean-square norm)

Let \(f\) be a function on \((-\pi, \pi)\) that is square-integrable, i.e.,
\[\int_{-\pi}^{\pi}|f(x)|^2dx < +\infty.\]
We define the mean-square norm or \(L^2\) norm \(\|f\|\) by
\[\|f\| = \sqrt{\int_{-\pi}^{\pi}|f(x)|^2dx}.\]

Remark. The \(L\) in \(L^2\) stands for ``Lebesgue.'' suggesting that we should use the Lebesgue integral rather than the Riemann integral. But the term ``\(L^2\)'' is so widespread that we use it, although we only use the Riemann integral. □

See also: Norm (Wikipedia)

Definition (Normed space)

A vector space on which a norm is defined is called a normed vector space or, simply, normed space. 

Definition (\(L^2\) inner product)

Let \(f, g\) be complex-valued, square-integrable functions on \((-\pi, \pi)\). We define the \(L^2\) inner product between \(f\) and \(g\) as
\[(f, g) = \int_{-\pi}^{\pi}f(x)\overline{g(x)}\,dx\]
where \(\overline{g(x)}\) is the complex conjugate of \(g(x)\). 

We need to verify that the integral on the right-hand side of the above equation does exist. In fact,
\[\begin{eqnarray*} \left|\int_{-\pi}^{\pi}f(x)\overline{g(x)}dx\right| &\leq& \int_{-\pi}^{\pi}|f(x)\overline{g(x)}|dx\\ &=& \int_{-\pi}^{\pi}|f(x)|\cdot |g(x)|dx \\ &\leq&\int_{-\pi}^{\pi}\frac{|f(x)|^2 + |g(x)|^2}{2}dx \\ &=& \frac{1}{2}\int_{-\pi}^{\pi}|f(x)|^2dx + \frac{1}{2}\int_{-\pi}^{\pi}|g(x)|^2dx\\ &<& +\infty \end{eqnarray*}\]

Remark. The \(L^2\) inner product corresponds to the scalar (dot) product in a vector space. □

In general, an inner product is defined as follows.

Definition (Inner product (general))

Let \(V\) be a vector space over the field \(K\). An inner product \((\cdot, \cdot): V\times V \to K\) is a map with the following properties for all vectors \(x,y,z\in V\) and all scalars \(\alpha, \beta \in K\):
  1. (Conjugate symmetry) \[(x,y) = \overline{(y,x)}.\]
  2. (Linearity in the first argument) \[(\alpha x + \beta y, z) = \alpha(x,z) + \beta(y,z)\]
  3. (Positive definiteness) If \(x \neq 0\), \[(x,x) > 0.\]

Example. Consider the vector space \(\mathbb{R}^n\). For \(x = (x_1, x_2, \cdots, x_n), y = (y_1, y_2, \cdots, y_n) \in \mathbb{R}^n\), the scalar product is defined as
\[(x,y) = \sum_{i=1}^{n}x_iy_i.\]
For each \(a \in \mathbb{R}\), its ``conjugate'' is the same \(a\): \(\overline{a} = a\). Thus, 
\[\begin{eqnarray}\overline{(y,x)} &=& (y,x)\\ &=& \sum_{i=1}^{n}y_ix_i\\ &=& \sum_{i=1}^{n}x_iy_i\\ &=& (x,y).\end{eqnarray}\]
For \(a,b\in\mathbb{R}\) and \(x,y,z \in \mathbb{R}^n\),
\[\begin{eqnarray}(ax+by,z) &=& \sum_{i=1}^{n}(ax_i+by_i)z_i\\ &=& a\sum_{i=1}^{n}x_iz_i + b\sum_{i=1}^{n}y_iz_i\\ &=& a(x,z) + b(y,z). \end{eqnarray}\]
If \(x \in \mathbb{R}^n\) is non-zero, at least one of its components, \(x_1, x_2, \cdots, x_n\), is non-zero. Thus,
\[(x,x) = \sum_{i=1}^{n}x_i^2 > 0.\]
Thus, the scalar product in \(\mathbb{R}^n\) is an inner product. □

Exercise. Show that \(L^2\) inner product satisfies the conditions of an inner product. □

Definition (Inner product space)

A vector space on which an inner product is defined is called an inner product space.

Example. \(\mathcal{R}_{2\pi}^2\) with the \(L^2\) inner product is an inner product space. □

Lemma (Cauchy-Schwarz inequality)

For \(f, g \in \mathcal{R}_{2\pi}^2\), we have
\[|(f, g)| \leq \|f\|\cdot \|g\|.\]
Proof. Exercise. ■

Exercise. Using the Cauchy-Schwarz inequality, show the triangle inequality:
\[\|f+g\| \leq \|f\| + \|g\|.\] □

Theorem 

Let \(f, f_1, f_2, g, g_1, g_2\) be square-integrable and \(\alpha\in \mathbb{C}\) be an arbitrary constant. The following hold:
\[\begin{eqnarray} \|f\|^2 &=& (f, f),\\ (f, g) &=& \overline{(g, f)},\\ (f_1 + f_2, g) &=& (f_1, g) + (f_2, g),\\ (f, g_1 + g_2) &=& (f, g_1) + (f, g_2),\\ (\alpha f, g) &=& \alpha(f, g),\\ (f, \alpha g) &=& \bar{\alpha}(f,g). \end{eqnarray}\]
Proof. Exercise. ■

Using the inner product, the Fourier coefficient of \(f\) can be written as
\[c_k = \frac{1}{2\pi}(f, e^{ikx}) ~~~ (k \in \mathbb{Z}).\]
The orthogonality of \(\{e^{inx}\}\) can be expressed as
\[(e^{inx}, e^{imx}) = 2\pi\delta_{n,m}\]
where \(\delta_{n,m}\) is Kronecker's delta.

Remark. If the function \(f\) has a period of \(2\pi\), the range of integration in
\[\|f\|^2 = \int_{-\pi}^{\pi}|f(x)|^2\,dx\]
can be anything, i.e., \(\int_a^{b}|f(x)|^2dx\) as long as \(b - a = 2\pi\). The same argument applies to the inner product.

Comments

Post a Comment

Popular posts from this blog

Birth process

-
Image
  Defining the birth process  Consider a colony of bacteria that never dies. We study the following process known as  the birth process , also known as  the Yule process . The colony starts with \(n_0\) cells at time \(t = 0\). Assume that the probability that any individual cell divides in the time interval \((t, t + \delta t)\) is proportional to \(\delta t\) for small \(\delta t\). Further assume that each cell division is independent of others. Let \(\lambda\) be the  birth rate.  The probability of a cell division for a population of \(n\) cells during \(\delta t\) is \(\lambda n \delta t\). We assume that the probability that two or more births take place in the time interval \(\delta t\) is \(o(\delta t)\). That is, it can be ignored. Consequently, the probability that no cell divides during \(\delta t\) is \(1 - \lambda n \delta t - o(\delta t)\). Note that this process is an example of the Markov chain with states \({n_0}, {n_0 + 1}, {n_0 + 2}...
Read more

Informal introduction to formal logic

-
Image
 In mathematics, we must prove (almost) everything and the proofs must be done logically and rigorously. Therefore, we need some understanding of basic logic. Here, I will informally explain some rudimentary formal logic. Definitions (Proposition): A proposition is a statement that is either true or false. "True" and "false" are called the truth values, and are often denoted \(\top\) and \(\bot\). Here is an example. "Dr. Akira teaches at UBD." is a statement that is either true or false (we understand the existence of Dr. Akira and UBD), hence a proposition. The following statement is also a proposition, although we don't know if it's true or false (yet): Any even number greater than or equal to 4 is equal to a sum of two primes. See also: Goldbach's conjecture   Next, we define several operations on propositions. Note that propositions combined with these operations are again propositions.   (Conjunction, logical "and"):  Let \(P\)...
Read more

Taylor's theorem for multivariate functions

-
Image
Taylor's theorem As we have seen in a previous post, the essence of differentiation is a linear approximation.  See also: Partial and total differentiation of multivariate functions Given a totally differentiable function \(z = f(x,y)\), we have in the neighbor of \(P = (a,b)\), \[f(x,y) = f(a,b) + f_x(a,b)(x - a) + f_y(a,b)(y - b) + o(\|X - P\|)\] where \(X = (x,y).\) By discarding the higher order terms \(o(\|X - P\|)\), we have the equation of the tangent plane at \(P\): \[z = f(a,b) + f_x(a,b)(x - a) + f_y(a,b)(y - b)\] that gives the linear approximation of \(z = f(x,y)\) around the point \(P\). But we may obtain better approximations by looking into the details of \(o(\|X - P\|)\), that is, by incorporating higher-order derivatives. Just as in the univariate case, we have a multivariate version of Taylor's theorem, which can be stated concisely and proved easily by using differential operators. See also: Differential operators Theorem (Taylor's theorem) Let \(f(x,y)...
Read more