Banach Spaces - Definitions and Examples, Continuous Linear Transformations, Banach Theorem, Natural Embedding of X in X''

1. Introduction to Normed Linear Spaces

Before we dive into Banach spaces, it's essential to understand the concept of a normed linear space. A linear space (or vector space) is a set of vectors that can be added together and multiplied by scalars, following certain axioms. A norm on a linear space is a function that assigns a strictly positive length or size to each vector, with the exception of the zero vector, which has a length of zero. This length is always a non-negative real number.

Formally, let V be a linear space over the field F (where F is either the set of real numbers ℝ or complex numbers ℂ). A function ||·|| : V → ℝ is called a norm if it satisfies the following properties for all vectors x, y in V and all scalars α in F:

• Non-negativity: ||x|| ≥ 0
• Definiteness: ||x|| = 0 if and only if x is the zero vector.
• Absolute homogeneity: ||αx|| = |α| ||x||
• Triangle inequality: ||x + y|| ≤ ||x|| + ||y||

A linear space V equipped with a norm ||·|| is called a normed linear space, denoted as (V, ||·||).

2. Definition of Banach Spaces

A Banach space is a special type of normed linear space that is "complete." Completeness is a crucial property in analysis, ensuring that Cauchy sequences converge to a limit within the space. This property is vital for the convergence of series and the existence of solutions to many types of equations.

Definition: A Banach space is a complete normed linear space.

What does "complete" mean here? In a complete metric space, every Cauchy sequence converges to a point within the space. A sequence {x_n} in a normed linear space is called a Cauchy sequence if for every ε > 0, there exists an integer N such that for all m, n > N, the distance between x_m and x_n, given by ||x_m - x_n||, is less than ε.

The completeness property ensures that if a sequence of vectors "looks like" it's converging (i.e., the terms get arbitrarily close to each other), then it actually does converge to a vector that exists within the space.

3. Examples of Banach Spaces

Understanding Banach spaces requires looking at concrete examples. These examples illustrate the diverse applications and fundamental nature of these spaces in mathematics and physics.

3.1. Euclidean Spaces ℝⁿ and ℂⁿ

The familiar n-dimensional real or complex vector spaces, ℝⁿ and ℂⁿ, are Banach spaces when equipped with standard norms. For instance, in ℝⁿ, the Euclidean norm (or L2 norm) of a vector x = (x₁, x₂, ..., x_n) is defined as:

||x||₂ = ( |x₁|² + |x₂|² + ... + |x_n|² )^1/2

Other common norms in ℝⁿ include the L1 norm (Manhattan distance) and the L∞ norm (maximum component):

||x||₁ = |x₁| + |x₂| + ... + |x_n|

||x||_∞ = max( |x₁|, |x₂|, ..., |x_n| )

All these normed spaces (ℝⁿ with ||·||_p for p=1, 2, ∞, and ℂⁿ similarly) are finite-dimensional and hence complete, making them Banach spaces.

3.2. Spaces of Continuous Functions C(K)

Let K be a compact Hausdorff topological space (e.g., a closed interval [a, b] in ℝ). The set of all continuous real-valued (or complex-valued) functions defined on K, denoted by C(K), forms a linear space. This space can be equipped with the supremum norm (or uniform norm), defined as:

||f||_∞ = sup_{t ∈ K} |f(t)|

This norm represents the maximum absolute value of the function over the compact set K. The space C(K) with this norm is a Banach space. The completeness ensures that if a sequence of continuous functions uniformly converges to a limit function, that limit function is also continuous.

3.3. Lp Spaces (1 ≤ p ≤ ∞)

For a measure space (X, Σ, μ) and a real number p such that 1 ≤ p < ∞, the Lp space, denoted by L^p(X, μ), consists of all measurable functions f: X → ℂ such that the integral of |f(x)|^p over X is finite. The norm is defined as:

||f||_p = ( ∫_X |f(x)|^p dμ(x) )^1/p

The space L^p(X, μ) is a Banach space for all 1 ≤ p < ∞. For p = ∞, the L^∞(X, μ) space consists of essentially bounded measurable functions, and its norm is the essential supremum of |f(x)|.

A particularly important case is the space L^p([a, b]) of Lebesgue integrable functions on the interval [a, b].

3.4. Sequence Spaces l^p

For 1 ≤ p < ∞, the space l^p consists of all infinite sequences x = (x₁, x₂, ...) of complex numbers such that the sum of |x_i|^p converges. The norm is defined as:

||x||_p = ( Σ_i=1^∞ |x_i|^p )^1/p

The space l^p is a Banach space. For p = ∞, the space l^∞ consists of all bounded sequences, with the norm being the supremum of the absolute values of the terms:

||x||_∞ = sup_i |x_i|

l^∞ is also a Banach space.

4. Continuous Linear Transformations

In functional analysis, we are often interested in mappings between normed linear spaces that preserve the linear structure and are "well-behaved" with respect to the norms. These are called continuous linear transformations (or bounded linear operators).

Definition: Let X and Y be normed linear spaces. A mapping T : X → Y is called a linear transformation if for all x₁, x₂ in X and all scalars α:

• T(x₁ + x₂) = T(x₁) + T(x₂)
• T(αx₁) = αT(x₁)

A linear transformation T : X → Y is said to be continuous if for every ε > 0, there exists a δ > 0 such that if ||x - y||_X < δ, then ||T(x) - T(y)||_Y < ε.

Equivalently, a linear transformation T is continuous if it maps convergent sequences to convergent sequences. That is, if x_n → x in X, then T(x_n) → T(x) in Y.

An extremely important result is that for linear transformations between normed spaces, continuity is equivalent to boundedness. A linear transformation T : X → Y is bounded if there exists a constant M ≥ 0 such that for all x in X:

||T(x)||_Y ≤ M ||x||_X

The smallest such constant M is called the norm of the operator T, denoted by ||T||. If T is a bounded linear transformation, its norm is defined as:

||T|| = sup_{x ≠ 0} ( ||T(x)||_Y / ||x||_X ) = sup_{||x||X = 1} ||T(x)||_Y

Theorem: A linear transformation T : X → Y between normed linear spaces X and Y is continuous if and only if it is bounded.

The set of all continuous linear transformations from a Banach space X to a Banach space Y forms a Banach space itself, denoted by B(X, Y), under the operator norm ||·||.

5. The Banach Theorem (Open Mapping Theorem)

The Banach theorem, more formally known as the Open Mapping Theorem, is a cornerstone of functional analysis. It provides a powerful condition under which a continuous linear transformation between Banach spaces is an open map and, consequently, its inverse is also continuous.

Statement of the Open Mapping Theorem: Let X and Y be Banach spaces, and let T : X → Y be a continuous linear transformation. If T is surjective (onto), then T is an open mapping. This means that for any open set U in X, the image T(U) is an open set in Y.

Consequences of the Open Mapping Theorem:

• If T : X → Y is a continuous linear bijection between Banach spaces X and Y, then its inverse T^-1 : Y → X is also continuous.
• This implies that if two Banach spaces are linearly homeomorphic (there exists a continuous linear bijection between them with a continuous inverse), then they are essentially the same from an analytical perspective.

Intuition behind the theorem: The theorem essentially says that if you have a "good" mapping (continuous, linear, surjective) between two complete spaces, it doesn't "collapse" dimensions in a way that would prevent its inverse from being well-behaved. The completeness of both spaces is crucial.

5.1. Proof Sketch (Conceptual)

The proof of the Open Mapping Theorem is non-trivial and often uses Baire's Category Theorem. A simplified conceptual outline:

1. Use Baire's Category Theorem to show that the image of a "large" ball in X under T contains a ball in Y. Specifically, show that T(B(0, 1)) contains a ball B(0, r) for some r > 0.
2. Extend this result from balls of radius 1 to balls of any radius, implying T maps open sets to open sets.
3. For the inverse mapping part, if T is a continuous linear bijection, then T maps the unit ball in X to a set that contains some ball B(0, r) in Y. By scaling and linearity, T maps the entire space X onto Y. The surjectivity ensures that every element in Y has a pre-image. The inverse T^-1 can be shown to be bounded by exploiting the fact that T maps some ball in X to a ball in Y.

6. The Natural Embedding of X in X''

Every normed linear space X can be embedded into its second dual space X'', which is the dual space of the dual space X*. The dual space X* of a normed linear space X is the space of all continuous linear functionals on X, equipped with the operator norm.

Definition of the Dual Space X*:

X* = { f : X → ℂ | f is a continuous linear functional on X }

The norm on X* is given by ||f||_X* = sup_||x||X=1 |f(x)|.

If X is a Banach space, then its dual space X* is also a Banach space.

Definition of the Second Dual Space X'':

X'' = (X*)* = { g : X* → ℂ | g is a continuous linear functional on X* }

X'' is also a Banach space if X* is a Banach space (which is true if X is a Banach space).

6.1. The Embedding Map J

We can define a map J : X → X'' as follows:

For each x ∈ X, define J(x) as a functional on X*. This functional, let's call it J(x) or sometimes denoted as $\hat{x}$, takes a functional f ∈ X* and maps it to a complex number:

(J(x))(f) = f(x) for all f ∈ X*.

This map J is called the natural embedding (or canonical embedding) of X into X''.

6.2. Properties of the Embedding Map J

1. J is linear: For x₁, x₂ ∈ X and α ∈ ℂ, J(αx₁ + x₂) = αJ(x₁) + J(x₂). This can be verified by applying the functionals to an arbitrary f ∈ X*.

2. J is an isometry: ||J(x)||_X'' = ||x||_X for all x ∈ X. Let's show this. The norm of J(x) in X'' is defined as:

||J(x)||_X'' = sup_{||f||X*=1, f ∈ X*} |(J(x))(f)| = sup_{||f||X*=1, f ∈ X*} |f(x)|

By the definition of the norm of f in X*, we know that |f(x)| ≤ ||f||_X* ||x||_X. Therefore, |f(x)| ≤ 1 ⋅ ||x||_X when ||f||_X* = 1.

So, sup_||f||X*=1 |f(x)| ≤ ||x||_X. This shows ||J(x)||_X'' ≤ ||x||_X.

To show the reverse inequality, the Hahn-Banach theorem guarantees that for any non-zero x ∈ X, there exists a functional f ∈ X* such that f(x) = ||x||_X and ||f||_X* = 1. For such an f, we have:

|(J(x))(f)| = |f(x)| = ||x||_X.

Therefore, ||J(x)||_X'' = sup_||f||X*=1 |f(x)| ≥ |f(x)| = ||x||_X.

Combining both inequalities, we get ||J(x)||_X'' = ||x||_X. This means J is an isometry.

3. J maps X into X''.

6.3. Reflexive Spaces

A Banach space X is called reflexive if the natural embedding map J : X → X'' is surjective. In other words, for a reflexive Banach space X, every continuous linear functional on its dual space X* can be represented by an element of X itself via the map J.

Theorem: All finite-dimensional normed linear spaces are reflexive. Also, L^p spaces for 1 < p < ∞ are reflexive. l^p spaces for 1 < p < ∞ are reflexive.

Counter-example: L¹([0, 1]) is not reflexive. Its dual space (L¹([0, 1]))* is isometrically isomorphic to L^∞([0, 1]), but the dual of L^∞([0, 1]) is not L¹([0, 1]).

The concept of reflexivity is important because it allows us to use the geometry of the space X'' to understand properties of X, especially in the context of optimization and analysis.

6.4. Significance of X''

The second dual space X'' often provides a "better" space to work in than X. For instance, weak convergence is defined using functionals from X*, but sometimes working with the bidual X'' can simplify arguments. Reflexive spaces have nicer properties concerning convergence (e.g., the Banach-Alaoglu theorem guarantees weak* compactness of the unit ball in X*, and for reflexive spaces, the unit ball in X is weakly compact).

The natural embedding J allows us to view X as a subspace of X''. If X is reflexive, X is essentially identified with X'' (up to an isometric isomorphism).