0% found this document useful (0 votes)
87 views11 pages

L Spaces: 5.1 Basic Theory

This document summarizes key concepts about L p spaces: - It defines L p spaces as generalizations of L1 spaces that contain measurable functions with finite p-norms for p between 0 and infinity. - It proves several properties of L p spaces including that they are vector spaces, the p-norm defines a norm for p ≥ 1, and L p spaces are complete normed spaces (Banach spaces). - It shows there are generally no inclusion relations between different L p spaces but some relations hold under certain conditions like if the measure space is finite.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
87 views11 pages

L Spaces: 5.1 Basic Theory

This document summarizes key concepts about L p spaces: - It defines L p spaces as generalizations of L1 spaces that contain measurable functions with finite p-norms for p between 0 and infinity. - It proves several properties of L p spaces including that they are vector spaces, the p-norm defines a norm for p ≥ 1, and L p spaces are complete normed spaces (Banach spaces). - It shows there are generally no inclusion relations between different L p spaces but some relations hold under certain conditions like if the measure space is finite.
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

Chapter 5

L p spaces

5.1 Basic theory


Throughout this chapter, we assume a fixed measure space (X, ⌃, µ).

Definition 5.1 Let f ∶ X → R be measurable. For p ∈ (0, ∞) we define


1�p
� f � p = �� � f � p dµ� .
X

(This expression can be infinite.)

Definition 5.2 Let p ∈ (0, ∞).


L p (µ) = { f ∶ X → R ∶ f is measurable and � f � p < ∞.}

L p (µ)-spaces are generalizations of L1 (µ)-spaces; the elements of L p (µ) are equiv-


alence classes of functions, which may di↵er on sets of zero measure.

Lemma 5.3 L p (µ) is a vector space.

Proof : Closure under scalar multiplication is obvious. Closure under addition


follows from the inequality

� f + g� p ≤ (2 max(� f �, �g�)) p ≤ 2 p (� f � p + �g� p ),


150 Chapter 5

hence
� f + g� pp ≤ 2 p �� f � pp + �g� pp �
—29h(2018)— n
We next show that as the notation indicates, � ⋅ � p is a norm on L p (µ) (for p ≥ 1).

Definition 5.4 Let p > 1. We denote its conjugate exponent (%$&/7 %8'() by
p∗ = p�(p − 1).

Note that
1 1
+ = 1.
p p∗

Lemma 5.5 (Young’s inequality) Let p > 1 and set q = p∗ . Then, for every a, b ∈
R:
�a� p �b�q
�ab� ≤ +
p q

Proof : Since (− log) is a convex function and since 1�p and 1�q sum up to one,
then, for every ↵, > 0:

1 1
− log � + � ≤ − log ↵ − log = − log(↵1�p ).
↵ 1�q
p q p q

It follows that:
+ ≥ ↵1�p
↵ 1�q
.
p q
Setting ↵ = �a� p and = �b�q we recover the desired result. n

Proposition 5.6 (Hölder inequality) Let p > 1 and set q = p∗ . For every measur-
able f, g ∶ X → R,
� f g�1 ≤ � f � p �g�q .
In particular, if f ∈ L p (µ) and q ∈ Lq (µ), then f g ∈ L1 (µ).
L p spaces 151

Proof : If either � f � p = 0 or �g�q = 0, then the result is trivial. Otherwise, using


Young’s inequality,

fg f g 1 � f �p 1 �g�q 1 1
� � =� dµ ≤ � p dµ + � q dµ = + = 1.
� f � p �g�q 1 X � f � p �g�q p X � f �p q X �g�q p q

Multiplying both sides by � f � p �g�q we obtain the desired result. n

Proposition 5.7 (Minkowski inequality) Let p ≥ 1. For every f, g ∈ L p (µ),


� f + g� p ≤ � f � p + �g� p .

Proof : The inequality is trivial for p = 1. For p > 1 set q = p∗ ; it follows from the
triangle inequality that

� f + g� p = � f + g� � f + g� p−1 ≤ � f � � f + g� p−1 + �g� � f + g� p−1 .

Integrating, and using Hölder’s inequality:

� f + g� pp ≤ � f � p � f + g� p−1
p + �g� p � f + g� p ,
p−1

where we used the defining property of q. This completes the proof. n

Corollary 5.8 For every p ≥ 1, � ⋅ � p is a norm on L p (µ).

Proof : Positivity and homogeneity are immediate; the triangle inequality is noth-
ing but Minkowski’s inequality. n

. Exercise 5.1 Show that � ⋅ � p is not a norm for 0 < p < 1.

Proposition 5.9 L p (µ) is a complete normed space, i.e., a Banach space ("(9/
+1").

. Exercise 5.2 Prove the completeness of L p (µ) in two steps:


152 Chapter 5

(a) Prove that it suffices to prove that for every sequence fn ∈ L p (µ),
∞ ∞
if � � fn � p < ∞ then � fn converges in L p (µ).
n=1 n=1

(b) Prove the statement by showing that ∑∞ n=1 � fn � p < ∞ implies that F = ∑n=1 fn exists a.e.;
∞ def

then show that F ∈ L p (µ) and that convergence is in L p (µ).

Proposition 5.10 For every p ≥ 1, the space of simple functions f = ∑ j a j Ej for


which µ(E j ) < ∞ for all j is dense in L p (µ).

Proof : Let f ∈ L p (µ). Then, there exist simple functions n ∈ SF+ (X) converging
monotonically to f + , and simple functions n ∈ SF+ (X) converging monotonically
to f − . Set fn = n − n . Then, fn → f pointwise, and

� fn � ≤ n + n ≤ f + + f − = � f �.

It follows that fn ∈ L p (µ), hence the E j in the representation of fn satisfy µ(E j ) <
∞.
Furthermore,
� fn − f � p ≤ (� fn � + � f �) p ≤ 2 p � f � p ,
i.e., � fn − f � p ∈ L1 (µ). It follows from dominated convergence that

lim � � fn − f � p dµ = 0.
n→∞ X

n
The spaces L p are generalizations of the space L1 (µ). In general, there is no
inclusion relations between those space:

Proposition 5.11 In general, for every p ≠ q there is no inclusion relation be-


tween L p (µ) and Lq (µ).
L p spaces 153

Proof : Consider the space ([1, ∞), B([1, ∞)), m) and set f (x) = 1�x. Then,

f ∈ L2 (m) but f ∈� L1 (m).



Consider the space ([−1, 1), B([−1, 1)), m) and set g(x) = 1� x. Then,

g ∈ L1 (m) but g ∈� L2 (m).

n
However, we have the following results:

Proposition 5.12 Let 1 < p < q < r . Then,


Lq (µ) ⊂ L p (µ) + Lr (µ).

That is, every f ∈ Lq (µ) can be represented as g + h, where g ∈ L p (µ) and h ∈


Lr (µ).

Proof : Let f ∈ Lq (µ) be given and let

E = {x ∶ � f (x)� > 1}.

Define g = f E and h = f Ec . Then, f = g + h, and

�g� p = � f � p E ≤ � f �q E and �h�r = � f �r Ec ≤ � f �q Ec ,

which proves that

�g� p ≤ � f �q and �h�r ≤ � f �q .

Proposition 5.13 Let 1 < p < q < r . Then,


L p (µ) ∩ Lr (µ) ⊂ Lq (µ).
154 Chapter 5

Proof : Let f ∈ L p (µ) ∩ Lr (µ). Since q�p > 1 and q�r < 1, there exists a ∈ (0, 1)
satisfying
q (1 − )q
+ = 1.
p r
Using Hölder’s inequality

� f �qq = � � f �q dµ = � � f � q � f �(1− )q
dµ = �� f � q � f �(1− )q
�1
X X

≤ �� f � q � p� q �� f � �r�(1− = � f � pq � f �r
(1− )q (1− )q
)q ,

i.e., � f �q ≤ � f � p � f �1−
r < ∞. n
If the measure space is finite, then the L p (µ) space satisfy the following inclusion
relation:

Proposition 5.14 If µ(X) < ∞ then 1 ≤ p < q implies that Lq (µ) ⊂ L p (µ).

Proof : Let f ∈ Lq (µ) and let

E = {x ∶ � f (x)� ≤ 1}.

Then,

�X � f � dµ = �E � f � dµ + �E c � f � dµ ≤ µ(E) + �E c � f � dµ < ∞.
p p p q

TA material 5.1 Define and treat the case of p = ∞.


. Exercise 5.3 Let (X, ⌃, µ) be a finite measure space, let f ∈ L p (µ) and let q = p∗ . Show
that
� f �1 ≤ (µ(X))1�q � f � p .

. Exercise 5.4 Consider the measure space ([0, 1], B([0, 1]), m) and let f ∈ L p (m) for some
p > 1. Let q = p . Prove that

1
lim 1�q � � f � dm = 0.
t↘0 t [0,t]
L p spaces 155

. Exercise 5.5 Let (X, ⌃, µ) be a finite measure space. Let fn , f ∈ L2 (µ), such that
� fn �2 ≤ M and lim fn = f a.e.
n→∞

(a) Prove that f ∈ L2 . (b) Use Egorov’s theorem to show that fn → f in L1 (µ).

. Exercise 5.6 Let µ be the counting measure on N and let 1 < p < q, ∞. Show that
� f � p ≤ � f �q .

5.2 Duality
Definition 5.15 Let p ≥ 1. A bounded linear functional (.&2( *9!*1*- -1&*781&5)
on L p (µ) is a linear mapping ∶ L p (µ) → R, satisfying
� � = sup{� ( f )� ∶ � f � p = 1} < ∞.
def

This space is called the space dual (*-!&$) to L p (µ) and is denoted (L p (µ))∗ .

The mapping � � � is the operator norm (;*9&)95&! %/9&1); you must have
seen in the past that it is indeed a norm on (L p (µ))∗ .
Also,

Proposition 5.16 Bounded linear functionals on L p (µ) are continuous.

Proof : Let fn → f in L p (µ) and let ∈ (L p (µ))∗ . Then,


fn − f
� ( fn ) − ( f )� = � ( fn − f )� = � � �� � fn − f � p ≤ � �� fn − f � p → 0.
� fn − f � p
n

Proposition 5.17 Let p > 1 and q = p∗ be its conjugate. Then, for every g ∈
Lq (µ), the map
g ∶ f � � f g dµ
X
is a bounded linear functional. Moreover,

� g � = �g�q .
156 Chapter 5

Proof : Clearly, q is a linear functional. Let � f � p = 1. By Hölder’s inequality,

� g ( f )� ≤ � � f g� dµ ≤ � f � p �g�q = �g�q ,
X

from which follows that � �g ≤ �g�q . Equality is obtained by taking,

�g�q−1
f = sgn(g) ,
�g�q−1
q

in which case
�g� p(q−1) �g�q
� f �p = � dµ = � q dµ = 1,
X �g�q
p(q−1) X �g�q

and
�g�q−1 �g�q
� g ( f )� = � sgn(g) g dµ = �X dµ = �g�q .
X �g�q−1
q �g�q−1
q

n
In other words, the mapping

Lq (µ) → (L p (µ))∗ ,

which maps g ∈ Lq (µ) into g ∈ (L p (µ))∗ is an isometric embedding (0&,*:


*9)/&'*!). The following seminal theorem shows that this map is in fact an isome-
try.

Theorem 5.18 (Riesz representation for L p -spaces) Let p > 1 and let q = p∗ .
Then
Lq (µ) � (L p (µ))∗
in the category of Banach spaces. That is, to every ∈ (L p (µ))∗ corresponds a
unique g ∈ Lq (µ) such that = g and � � = �g�q .

Proof : We will only prove the theorem for the case of a finite measure space. It
remains to prove that to every corresponds a unique g such that = g .
Step 1: uniqueness: If g = h, then for every f ∈ L p (µ),

�X f g dµ = �X f h dµ,
L p spaces 157

hence for every f ∈ L p (µ),

0 = � f (g − h) dµ.
X

Suppose that g ≠ h. Then, one of the sets


{x ∶ g(x) > f (x)} or {x ∶ g(x) < f (x)}.
has positive measure. Without loss of generality, we may assume that this is the
first. Then, for some n,
1
En = �x ∶ g(x) − h(x) > �
n
has positive measure. Setting f = En we obtain a contradiction.
Step 2: construct a signed-measure ⌫: Let ∈ (L p (µ))∗ be given. Since µ is
finite, E ∈ L p (µ) for every measurable set E ∈ ⌃; define
⌫(E) = ( E ).

We will show that ⌫ is a signed measure:

(a) Since � = 0 as an element of L p (µ), and is linear,


⌫(�) = ( E) = (0) = 0.

(b) Let

E = � En ,
n=1
then n
lim � Ek = E pointwise,
n→∞
k=1
and
n ∞ 1�p ∞ 1�p
� E −� Ek � =� � Ek � = �� dµ� = �µ � � Ek �� .
k=1 k=n+1 �∞
k=n+1 E k k=n+1
p p

Letting n → ∞, using the upper-semicontinuity of µ and the fact that �∞


n=1 �k=n+1 E k =

�,
n ∞ 1�p ∞ ∞ 1�p
lim � E −� Ek � = lim �µ � � Ek �� = �µ � � � Ek �� = 0,
n→∞ n→∞
k=1 p k=n+1 n=1 k=n+1
158 Chapter 5

i.e.,
n
lim � Ek = E in L p (µ).
n→∞
k=1
—35h(2017)—
Since is linear and continuous,
∞ ∞
⌫(E) = ( E) =� ( En ) = � ⌫(En ).
n=1 n=1

proving that ⌫ is countably-additive, i.e., it is a signed measure.

Step 3: Show that ⌫ � µ and apply Radon-Nikodym to obtain a function g:


Suppose that µ(E) = 0. Then, E = 0 (as an element of L p (µ)), hence

⌫(E) = ( E) = 0.

It follows from the Radon-Nikodym theorem that there exists an integrable func-
tion g ∈ L1 (µ), such that
⌫(E) = � g dµ,
E
which amounts to
( E) =� g E dµ.
X
By linearity,

( ) = � g dµ for all simple functions ∶ X → R.


X

By definition of the norm � �,

� ( )� = �� g dµ� ≤ � �� � p for all simple functions ∶ X → R.


X

Step 4: Prove that g ∈ Lq (µ): (Recall that Lq (µ) ⊂ L1 (µ), but not the other way
around.) Let gn be a sequence of simple functions converging to g pointwise, such
that �gn � ≤ �g�. By Fatou’s lemma for �gn � → �g�,

�g�q ≤ lim inf �gn �q .


n→∞

Set
�gn �q−1
fn = sgn(g) ,
�gn �q−1
q
L p spaces 159

which is a sequence of simple functions satisfying � fn � p = 1. Then,

1
�gn �q = � �gn �q dµ
�gn �q−1 X

= � � fn gn � dµ
X

≤ � � fn g� dµ
X

= � fn g dµ
X
= ( fn ) ≤ � �.

In the passage to the second line we used the definition of fn ; in the passage to the
third line we used the fact that �gn � ≤ �g�; in the passage to the fourth line we used
the fact that fn g > 0; in the passage to the fifth line we used the characterization
of for simple functions; in the passage to the sixth line we used the definition of
the operator norm and the fact that � fn � p = 1.
It follows that
�g�q ≤ lim inf �gn �q ≤ � �.
n→∞

Step 5: Prove that ∈ g : Finally, let f ∈ L p (µ). Since the space of simple
functions is dense in L p (µ) and is continuous, for n → f in L p (µ),

�� g f dµ − ( f )� = lim �� g f dµ − ( n )�
X n→∞ X

= lim �� g f dµ − � g n dµ�
n→∞ X X

= lim �� g( f − n ) dµ�
n→∞ X
≤ lim �g�q � f − n�p = 0.
n→∞

This completes the proof for the case where µ is a finite measure. n —36h(2017)—

You might also like