• 沒有找到結果。

A subalgebra of C(X) is a vector subspace A such that for any f, g ∈ A, f · g ∈ A

N/A
N/A
Protected

Academic year: 2022

Share "A subalgebra of C(X) is a vector subspace A such that for any f, g ∈ A, f · g ∈ A"

Copied!
3
0
0

加載中.... (立即查看全文)

全文

(1)

1. Stone-Weierstrass Theorem

Let X be a compact metric space and C(X) be the space of continuous real-valued functions on X. On C(X), we define

kf k= sup

x∈X

|f (x)|.

Then the normed space (C(X), k · k) is a Banach space over R.

A subalgebra of C(X) is a vector subspace A such that for any f, g ∈ A, f · g ∈ A. We say that a subalgebra A separates points if for any x, y ∈ X, there exists f ∈ A so that f (x) 6= f (y). We say that A is unital if A contains 1.

Theorem 1.1. (Stone-Weierstrass Theorem) Suppose A is a unital subalgebra of C(X) such that A separates points of X. Then A is dense in C(X), i.e. the closure A of A is the whole space C(X).

Proof. The proof will be omitted here. 

Corollary 1.1. (Weierstrass second approximation theorem) Let A be the vector subspace of C([0, 2π]) spanned by {1, cos nx, sin nx : n ≥ 1}. Then A is a unital subalgebra of A which separates points of [0, 2π] and hence A is dense in C(X).

Proof. Elements of A are are of the form

t(x) = a0

2 +

N

X

k=1

{akcos kθ + bksin kθ},

where a0, a1, · · · , aN and b1, · · · , bN are real numbers. To show that A is a subalgebra of C([0, 2π]), we need to show that t(x)s(x) ∈ A for all t(x), s(x) ∈ A. To do this, we use the following formulas:

(1) 2 cos kθ cos jθ = cos(k + j)θ + cos(k − j)θ, (2) 2 sin kθ sin jθ = cos(k − j)θ − cos(k + j)θ, (3) 2 sin kθ cos jθ = sin(k + j)θ + sin(k − j)θ.

Hence t(x)s(x) is also a linear combination of {1, cos nx, sin nx : n ≥ 1}. It is not hard to see that A separates points. By Stone-Weierstrass theorem, A is dense in C([0, 2π]).  Lemma 1.1. Let {fn} be a sequence of real-valued continuous functions on [a, b] which converges uniformly to a continuous function f on [a, b]. Then

n→∞lim Z b

a

fn(x)dx = Z b

a

f (x)dx.

Proof. Since {fn} converges to f uniformly on [a, b], for any  > 0, there exists N > 0 so that for any n ≥ N, and any x ∈ [a, b]

|fn(x) − f (x)| ≤ kfn− f k< .

1

(2)

2

Hence if n ≥ N,

Z b a

fn(x)dx − Z b

a

f (x)dx

=

Z b a

(fn(x) − f (x))dx

≤ Z b

a

|fn(x) − f (x)|dx

≤ Z b

a

kfn− f kdx

< (b − a).

By definition, we prove our assertion. 

Corollary 1.2. Let {fn} be a sequence of continuous functions on [a, b] so that the infinite series

X

n=1

fn(x) converges uniformly on [a, b]. Then Z b

a

X

n=1

fn(x)

! dx =

X

n=1

Z b

a

fn(x)dx.

Proof. Let us consider the n-th partial sum sn(x) =

n

X

k=1

fk(x). Then {sn} converges to a

function s =

X

n=1

fn in C([a, b]). By the previous lemma,

n→∞lim Z b

a

sn(x)dx = Z b

a

s(x)dx.

 Suppose ψ ∈ C([0, 2π]) periodic of period 2π such that the following series converges uniformly on [0, 2π] :

ψ(θ) = a0 2 +

X

n=1

{ancos nx + bnsin nx}.

By the uniform convergence, we can interchange summation and integral Z

0

ψ(θ)dθ = Z

0

ψ(θ)dθ +

X

n=1

{an Z

0

cos nθdθ + bn

Z 0

ψ(θ) sin nθdθ}.

Hence we obtain

a0 = 1 π

Z 0

ψ(θ)dθ.

Multiplying ψ by cos kθ, we obtain ψ(θ) cos kθ = a0

2 cos kθ +

X

n=1

{ancos nθ cos kθ + bnsin nθ cos kθ}.

Integrating this infinite series, by the uniform convergence of the infinite series, we have Z

0

ψ(θ) cos kθdθ = Z

0

a0

2 cos kθdθ +

X

n=1

Z 0

{ancos nθ cos kθ + bnsin nθ cos kθ}dθ.

(3)

3

Similarly, multiplying ψ by sin kθ and integrating it, we obtain Z

0

ψ(θ) sin kθdθ = Z

0

a0

2 sin kθdθ +

X

n=1

Z 0

{ancos nθ sin kθ + bnsin nθ sin kθ}dθ.

Using the trigonometric identity, one can show that for any n, k, (1)

Z 0

cos nθ cos kθdθ = πδnk, (2)

Z 0

sin nθ sin kθdθ = πδnk, (3)

Z 0

cos nθ sin kθdθ = 0.

These imply that for n ≥ 1, an= 1

π Z

0

ψ(θ) cos nθdθ, bn= 1 π

Z 0

ψ(θ) sin nθdθ.

參考文獻

相關文件

By the Lebesgue’s theorem we know that if f is (Riemann) integrable on A, then the discontinuities of f

Determine whether the following sequence of functions {f n } converges pointwise on the indicated interval2. (a) Suppose that {f n } is a sequence of bounded (not

Generalization Theorem Let f be integrable on K = [a, b] × [c, d] to R and suppose that for each y ∈ [c, d], the function x 7→ f (x, y) of [a, b] into R is continuous except

Date: Monday, Jun 17, 2002 Instructor: Shu-Yen Pan1. No credit will be given for an answer

Then g is defined on [a, b], satifies (11), and is continuous on [a, b] by the Sequential Characterization of Limits.. Thus, f

The set of all adherent points of A is denoted by A called the closure of A.. When A is empty, the statement

[r]

1、曾擔任以國家、重要城市為名,至少以二個版面以上刊登國際 新聞,且發行對象以全國或全球讀者為目標之平面媒體或通訊 社(例如:《美國新聞與世界報導》(U.S. News