MIT OpenCourseWare http://ocw.mit.edu
18.102 Introduction to Functional Analysis
Spring 2009
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
4
LECTURE NOTES FOR 18.102, SPRING 2009
Lecture 1. Tuesday, 3 Feb.
Linear spaces, metric spaces, normed spaces. Banach spaces. Examples – Eu clidean spaces, continuous functions on a closed interval – C ^{0} ([0, 1]) with supremum norm. The (Riemannian) L ^{1} norm, outline that this is not complete on C ^{0} ([0, 1]). Brief description of l _{2} – This is Hilbert space, but not explained.
What is it all for? Main aims: To be able to carry out ‘standard’ constructions in (linear) functional analysis:
Abstract Hilbert space – one in each dimension Concrete Hilbert space – Many, such as L ^{2} ([0, 1]) Example of a theorem: The Dirichlet problem. Let V : [0, 1] −→ R be a realvalued function. We are interested in ‘oscillating modes’ on the interval; something like this arises in quantum mechanics for instance. Namely we want to know about functions u(x) – twice continuously diﬀerentiable on [0, 1] which satisfy the diﬀerential equation
(1.1)
d ^{2} u
− _{d}_{x} _{2} (x) + V (x)u(x) = λu(x)
where λ is an ‘unknown’ constant – that is we want to know which λ’s can occur. Well, of course all λ’s can occur with u ≡ 0 but this is the ‘trivial solution’ which will always be there for such an equation. What other solutions are there? Well, there is an inﬁnite sequence of λ’s for which there is a non trivial solution of (1.1) λ _{j} ∈ R – they are all real no nonreal complex λ’s can occur. For each of these there is at least one (and maybe more) ‘independent’ solution u _{j} . We can say a lot more about everything here but one main aim of this course is to get at least to this point. Now, in fact (1.1) is just the eigenvalue equation. What we are dealing with here is an ‘inﬁnite matrix’. This is not obvious, and in fact is not a very good way of looking at things (there was such a matrix approach to quantum mechanics in the early days but it was replaced by the sort of ‘operator’ theory on Hilbert space that we will use here.) Still, we are in some sense dealing with inﬁnite dimensional matrices. One of the crucial diﬀerences between inﬁnite and ﬁnite dimensional settings is that topology is encountered. This is enshrined here in the notion of a normed linear space Linear space: Should I break out the axioms? It is a space V in which we can add elements and multiply by scalars with rules very similar to the basic examples of R ^{n} or C ^{n} . Note that for us the ‘scalars’ are either the real numbers of the complex numbers – usually the latter. Let’s be neutral and denote by K either R or C but of course consistently. Then our set V – the set of vectors with which we will deal, comes with two ‘laws’. These are maps
(1.2)
which we denote not by +(v, w) and ·(s, v) but by v + w and sv. Then we impose the axioms of a vector space – look them up! These are commutative group axioms for + axioms for the action of K and the distributive law. Our examples:
The ‘trivial’ case of a ﬁnite dimensional vector space. The l ^{p} spaces. These are spaces of sequences – the ﬁrst problem set is all about them. Thus l ^{2} – which is a Hilbert space – consists of all the sequences
+ : V × V
−→ V,
· : K × V
−→ V.
LECTURE NOTES FOR 18.102, SPRING 2009
5
(1.3)
a : N −→ C also denoted {a _{j} }
∞
j=1
∞
where a _{j} = a(j), such that
� a _{j}  ^{2} < ∞.
j=1
Seriously nontrivial examples such as C([0, 1]) the space of continuous func tions on [0, 1] (say with complex values). All these vector spaces carry norms, and that is what we want to talk about for most of the semester.
Deﬁnition 1. A norm on a vector space V is a function
(1.4)
�
·
�
:
V
−→ [0, ∞)
which satisﬁes the three properties:
(1) 
�v� = 0 iﬀ v = 0. 
(2) 
�tv� = t�v� for all t ∈ K and v ∈ V. 
(3) Triangle inequality. �v + w� ≤ �v� + �w� for all v, w ∈ V.
Then show that d(v, w) = �v − w� is a metric on V. It follows that all the notions from 18.100 come into play – open sets, balls, closed sets, convergence of sequences, compact sets, connected sets, complete metric spaces. We will use them all!
Deﬁnition 2. A normed space which is complete with respect to the induced metric is called a Banach space. I then discussed the supremum norm on C([0, 1]);
(1.5)
�u�
∞
=
x∈[0,1] 
sup
u(x) .

I said, but did not prove, that it this gives a Banach space. I also disussed the L ^{1} norm on C([0, 1]) :
(1.6)
�u� _{L} 1 =
�
0
1
u(x)dx
and indicated why it is not complete. This is basically why we need to study the Lebesgue integral. One might ask ‘is the space of Riemann integrable functions on [0, 1] complete with respect to the L ^{1} norm’ – indeed someone did. The answer is that �u� _{L} 1 is not even a norm on the space of Riemann integable functions. Namely it is only a ‘seminorm’ – the second two conditions are ﬁne but there are nonzero functions with integral zero. So, the question is not quite precise, which is why I did not talk about it. One the other hand one can ﬁddle with the space (take a quotient) so that one gets a norm and then it is not complete. So the morally correct answer is NO, it is not a Banach space and this is true for a better (or worse) reason than it simply not being a normed space in the ﬁrst place! We will get to this.
6
LECTURE NOTES FOR 18.102, SPRING 2009
Problem set 1, Due 11AM Tuesday 10 Feb.
Full marks will be given to anyone who makes a good faith attempt to answer each question. The ﬁrst four problems concern the ‘little L p’ spaces l ^{p} . Note that you have the choice of doing everything for p = 2 or for all 1 ≤ p < ∞. Problem 1.1 Write out a proof (you can steal it from one of many places but at least write it out in your own hand) either for p = 2 or for each p with 1 ≤ p < ∞ that
l ^{p} = {a : N −→ C;
_{∞}
�
j=1
a _{j}  ^{p} < ∞,
is a normed space with the norm
�a� _{p} =
⎛
⎝
∞
�
j=1
a _{j} 
p
⎞
⎠
1
p
.
a _{j} = a(j)}
This means writing out the proof that this is a linear space and that the three
conditions required of a norm hold. Problem 1.2 The ‘tricky’ part in Problem 1.1 is the triangle inequality. Suppose you knew – meaning I tell you – that for each N
⎛
⎝
N
�
j=1
a _{j} 
p
⎞
⎠
1
p
is a norm on C ^{N}
would that help? Problem 1.3 Prove directly that each l ^{p} as deﬁned in Problem 1.1 – or just l ^{2} – is complete, i.e. it is a Banach space. At the risk of oﬀending some, let me say that this means showing that each Cauchy sequence converges. The problem here is to ﬁnd the limit of a given Cauchy sequence. Show that for each N the sequence in C ^{N} obtained by truncating each of the elements at point N is Cauchy with respect to the norm in Problem 1.2 on C ^{N} . Show that this is the same as being Cauchy in C ^{N} in the usual sense (if you are doing p = 2 it is already the usual sense) and hence, this cutoﬀ sequence converges. Use this to ﬁnd a putative limit of the Cauchy sequence and then check that it works. Problem 1.4 Consider the ‘unit sphere’ in l ^{p} – where if you want you can set p = 2. This is the set of vectors of length 1 :
S = {a ∈ l ^{p} ;�a� _{p} = 1}.
(1) Show that S is closed. (2) Recall the sequential (so not the open covering deﬁnition) characterization of compactness of a set in a metric space (e.g. by checking in Rudin). (3) Show that S is not compact by considering the sequence in l ^{p} with kth element the sequence which is all zeros except for a 1 in the kth slot. Note that the main problem is not to get yourself confused about sequences of sequences! Problem 1.5 Show that the norm on any normed space is continuous.
MIT OpenCourseWare http://ocw.mit.edu
18.102 Introduction to Functional Analysis
Spring 2009
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
LECTURE NOTES FOR 18.102, SPRING 2009
7
Lecture 2. Thursday, 5 Feb.
Plan: Linear maps between normed spaces are continuous iﬀ they are bounded. The best bound gives a norm. If the second space is Banach the space of linear operators is Banach. Corollary – the dual space of a normed space is a Banach space. Examples: Integral operators on C ^{0} ([0, 1]) with respect to supremum or L ^{1} norms. Diﬀerentiation as an operator from C ^{1} ([0, 1]) to C ^{0} ([0, 1]) In practice: I did not get to the part about diﬀerentiation. Reading:
(1) Wilde: Chapter 2 to 2.7 (2) Chen: First part of Chapter 6 and of Chapter 7. (3) Ward: Chapter 3, ﬁrst 2 sections.
MIT OpenCourseWare http://ocw.mit.edu
18.102 Introduction to Functional Analysis
Spring 2009
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
8
LECTURE NOTES FOR 18.102, SPRING 2009
Lecture 3. Tuesday, 10 Feb.
Recalled the proof from last time that the bounded operators from a normed space into a Banach space form a Banach space – mainly to suggest that it is not so hard to remember how such a proof goes. Then proved that a normed space is Banach iﬀ every ‘absolutely summable’ series is convergent. Absolute summability means that the sum of the norms is ﬁnite. Then did most of the proof that every normed space can be completed to a Banach space using this notion of absolutely summable sequences. The last part – and a guide to how to attempt the part of the proof that is the ﬁrst question on the next homework. The proof of the result about completeness from the early part of the leture is in (1) Wilde: Proposition 1.6 (2) Chen: I didn’t ﬁnd it. (3) Ward: Lemma 2.1 (easy way only) Here is a slightly abbreviated version of what I did in lecture today on the completion of a normed space. The very last part I asked you to ﬁnish as the ﬁrst part of the second problem set, not due until February 24 due to the vagaries of the MIT calendar (but up later today). This problem may seem rather heavy sledding but if you can work through it all you will understand, before we get to it, the main sorts of arguments needed to prove most of the integrability results we will encounter later. Let V be a normed space with norm � · � _{V} . A completion of V is a Banach space B with the following properties: (1) There is an injective (11)linear map I : V −→ B (2) The norms satisfy
(3.1)
(3) The range I(V ) ⊂ B is dense in B. Notice that if V is itself a Banach space then we can take B = V with I the identity map. So, the main result is:
Theorem 1. Each normed space has a completion.
�I(v)� _{B} = �v� _{V}
∀ v ∈ V.
‘Proof ’ (the last bit is left to you). First we introduce the rather large space
(3.2)
�
V =
�
{u _{k} } _{k}_{=}_{1} ; u _{k} ∈ V and
∞
∞
�
k=1
�u _{k} � < ∞
�
the elements of which I called the absolutely summable series in V.
�
Now, I showed in the earlier result that each element of V is a Cauchy sequence
N
– meaning the corresponding sequence of partial sums v _{N} =
� u _{k} is Cauchy if
k=1
�
{u _{k} } is absolutely summable. Now V is a linear space, where we add sequences,
and multiply by constants, by doing the operations on each component:
(3.3) t _{1} {u _{k} } + t _{2} {u ^{�} _{k} } = {t _{1} u _{k} + t _{2} u _{k} ^{�} }.
This always gives an absolutely summable series by the triangle inequality:
(3.4)
�
�t _{1} u _{k} + t _{2} u ^{�} _{k} � ≤ t _{1} 
�
�u _{k} � + t _{2} 
�
�u _{k} ^{�} �.
k
k
k
LECTURE NOTES FOR 18.102, SPRING 2009
9
�
Within V consider the linear subspace
(3.5)
S
=
�
{u _{k} }; ^{�} �u _{k} � < ∞,
k
�
k
u _{k} = 0
�
of those which converge to 0. As always for a linear subspace of a linear space we can form the quotient
B = V /S
the elements of which are the ‘cosets’ of the form {u _{k} } + S ⊂ V where {u
(3.6)
_{k} } ∈ V .
We proceed to check the following properties of this B. (1) A norm on B is deﬁned by
�
n
(3.7)
�
� �
�b� _{B} =
n→∞ ^{�}
lim
u _{k} �,
{u _{k} } ∈ b.
k=1
(2) The original space V is imbedded in B by
(3.8) V � v �−→ I(v) = {u _{k} } + S, u _{1} = v, u _{k} = 0 ∀ k > 1
and the norm satisﬁes (3.1). (3) I(V ) ⊂ B is dense. (4) B is a Banach space with the norm (3.7). So, ﬁrst that (3.7) is a norm. The limit on the right does exist since the limit of the norm of a Cauchy sequence always exists – namely the sequence of norms
is itself Cauchy but now in R. Moreover, adding an element of S to {u _{k} } does not change the norm of the sequence of partial sums, since the addtional term tends to zero in norm. Thus �b� _{B} is welldeﬁned for each element b ∈ B and �b� _{B} = 0 means exactly that the sequence {u _{k} } used to deﬁne it tends to 0 in norm, hence is in S hence b = 0. The other two properties of norm are reasonably clear, since
�
if b, b ^{�} ∈ B are represented by {u _{k} }, {u ^{�} _{k} } in V then tb and b + b ^{�} are reprented by
{tu _{k} } and {u _{k} + u _{k} ^{�} } and
_{(}_{3}_{.}_{9}_{)}
n→∞ ^{�}
lim
n
�
k=1
tu _{k} � =

t
^{} n→∞ ^{�}
lim
n
�
k=1
u _{k} �,
n→∞ ^{�}
lim
n
�
k=1
(u _{k} + u ^{�} )� = A =⇒
k
for � > 0 ∃ N s.t.
∀ n ≥ N,
A − � ≤ �
n
�
k=1
u _{k} � + �
A − � ≤ �
n
�
k=1
(u _{k} + u _{k} ^{�} )� =⇒
n
�
k=1
u
�
k
)� ∀ n ≥ N =⇒
A − � ≤ �b� _{B} + �b ^{�} � _{B} ∀ � > 0 =⇒
�b + b ^{�} � _{B} ≤ �b� _{B} + �b ^{�} � _{B} .
Now the norm of the element I(v) = v, 0, 0, · · · , is the limit of the norms of the sequence of partial sums and hence is �v� _{V} so �I(v)� _{B} = �v� _{V} and I(v) = 0 therefore implies v = 0 and hence I is also injective. So, we need to check that B is complete, and also that I(V ) is dense. Here is an extended discussion of the diﬃculty – of course maybe you can see it directly
10
LECTURE NOTES FOR 18.102, SPRING 2009
yourself (or have a better scheme). Note that I want you to write out your own version of it carefully for the next problem set. Okay, what does it mean for B to be a Banach space, well as we saw in class
today it means that every absolutely summable series in B is convergent. Such a
series {b _{n} } is given by b _{n} = {u ^{(}^{n}^{)}
_{k} } + S where {u _{k} ^{(}^{n}^{)} } ∈ V and the summability
�
condition is that 

(3.10) 
∞ > 
�
n
�b _{n} � _{B} =
�
n
N →∞ ^{�}
lim
N
�
k=1
u
(n)
k
� _{V} .
So, we want to show that
limit b. It is supposed to be given by an absolutely summable series. The ‘problem’ is that this series should look like ^{�}^{�} u _{k} ^{n}^{)} in some sense – because it is supposed
b _{n} = b converges, and to do so we need to ﬁnd the
�
n
(
n
k
to represent the sum of the b _{n} ’s. Now, it would be very nice if we had the estimate
(3.11)
��
�u
^{n}^{)} � _{V}
(
_{k}
< ∞
n k
since this should allow us to break up the double sum in some nice way so as to get an absolutely summable series out of the whole thing. The trouble is that (3.11) need not hold. We know that each of the sums over k – for given n – converges, but not the sum of the sums. All we know here is that the sum of the ‘limits of the norms’ in (3.10) converges. So, that is the problem! One way to see the solution is to note that we do not have to choose the original {u ^{(}^{n}^{)} } to ‘represent’ b _{n} – we can add to it any element of S. One idea is to rearrange the u _{k} ^{n}^{)} – I am thinking here of ﬁxed n – so that it ‘converges even faster.’ Given � > 0 we can choose N _{1} so that for all N ≥ N _{1} ,
(3.12)
k
(
�
�
u _{k} ^{N}^{)} � _{V} − �b _{n} � _{B}  ≤ �,
(
�
�u _{k} ^{n}^{)} � _{V}
(
≤ �.
k≤N
k≥N
Then in fact we can choose successive N _{j} < N _{j}_{−}_{1} (remember that little n is ﬁxed here) so that
(3.13)
�
�
u _{k} ^{N}^{)} � _{V} − �b _{n} � _{B}  ≤ 2 ^{−}^{j} �,
(
�
�u
^{n}^{)} � _{V} ≤ 2 ^{−}^{j} �.
(
_{k}
k≤N _{j}
k≥N _{j}
N
1
Now, ‘resum the series’ deﬁning instead v _{1} ^{(}^{n}^{)} = ^{�} u
k=1
( n) _{,} _{v} j (n) _{=}
k
this setting � = 2 ^{−}^{n} for the nth series. Check that now
(3.14)
��
�v _{k} ^{(}^{n}^{)} � _{V} < ∞.
n k
N
j
�
k=N j−1
u _{k} ^{n}^{)} and do
(
Of course, you should also check that b _{n} = {v _{k} ^{(}^{n}^{)} } + S so that these new summable series work just as well as the old ones. After this ﬁddling you can now try to ﬁnd a limit for the sequence as
w _{k} =
(3.15)
b = {w _{k} } + S,
�
v ^{(}^{p}^{)} ∈ V.
l
l+p=k
So, you need to check that this {w _{k} } is absolutely summable in V and that b _{n} → b as n → ∞. I may add some more to discussion of completeness if needed.
LECTURE NOTES FOR 18.102, SPRING 2009
11
Finally then there is the question of showing that I(V ) is dense in B. You can do this using the same idea as above – in fact it might be better to do it ﬁrst. Given an element b ∈ B we need to ﬁnd elements in V, v _{k} such that �I(v _{k} ) − b� _{B} → 0 as
k → ∞. Take an absolutely summable series u _{k} representing b and take v _{j} =
u k
where the N _{j} ’s are constructed as above and check that I(v _{j} )
(3.16)
N j
�
k=1
→ b by computing
�I(v _{j} ) − b� _{B} =
p→∞ ^{�}
lim
�
p>N _{j}
u _{p} � _{V}
≤
�
p>N _{j}
�u _{p} � _{V} .
�
12
LECTURE NOTES FOR 18.102, SPRING 2009
Problem set 2, Due 11AM Tuesday 24 Feb.
I was originally going to make this problem set longer, since there is a missing Tuesday. However, I would prefer you to concentrate on getting all four of these questions really right! Problem 2.1 Finish the proof of the completeness of the space B constructed in lecture on February 10. The description of that construction can be found in the notes to Lecture 3 as well as an indication of one way to proceed. Problem 2.2 Let’s consider an example of an absolutely summable sequence of step functions. For the interval [0, 1) (remember there is a strong preference for leftclosed but rightopen intervals for the moment) consider a variant of the construction of the standard Cantor subset based on 3 proceeding in steps. Thus, remove the ‘central interval [1/3, 2/3). This leave C _{1} = [0, 1/3) ∪ [2/3, 1). Then remove the central interval from each of the remaining two intervals to get C _{2} = [0, 1/9) ∪ [2/9, 1/3) ∪ [2/3, 7/9) ∪ [8/9, 1). Carry on in this way to deﬁne successive sets C _{k} ⊂ C _{k}_{−}_{1} , each consisting of a ﬁnite union of semiopen intervals. Now, consider the series of step functions f _{k} where f _{k} (x) = 1 on C _{k} and 0 otherwise.
(1) Check that this is an absolutely summable series.
(2) For which x ∈ [0, 1) does
�
k
f _{k} (x) converge?
(3) Describe a function on [0, 1) which is shown to be Lebesgue integrable (as deﬁned in Lecture 4) by the existence of this series and compute its Lebesgue integral. (4) Is this function Riemann integrable (this is easy, not hard, if you check the deﬁnition of Riemann integrability)? (5) Finally consider the function g which is equal to one on the union of all the intervals which are removed in the construction and zero elsewhere. Show that g is Lebesgue integrable and compute its integral.
Problem 2.3 The covering lemma for R ^{2} . By a rectangle we will mean a set of the form [a _{1} , b _{1} ) × [a _{2} , b _{2} ) in R ^{2} . The area of a rectangle is (b _{1} − a _{1} ) × (b _{2} − a _{2} ).
(1) We may subdivide a rectangle by subdividing either of the intervals – re placing [a _{1} , b _{1} ) by [a _{1} , c _{1} ) ∪ [c _{1} , b _{1} ). Show that the sum of the areas of rectangles made by any repeated subdivision is always the same as that of the original. (2) Suppose that a ﬁnite collection of disjoint rectangles has union a rectangle (always in this same halfopen sense). Show, and I really mean prove, that the sum of the areas is the area of the whole rectange. Hint: proceed by subdivision. (3) Now show that for any countable collection of disjoint rectangles contained in a given rectange the sum of the areas is less than or equal to that of the containing rectangle. (4) Show that if a ﬁnite collection of rectangles has union containing a given rectange then the sum of the areas of the rectangles is at least as large of that of the rectangle contained in the union. (5) Prove the extension of the preceeding result to a countable collection of rectangles with union containing a given rectangle.
Problem 2.4
LECTURE NOTES FOR 18.102, SPRING 2009
13
(1) Show that any continuous function on [0, 1] is the uniform limit on [0, 1) of a sequence of step functions. Hint: Reduce to the real case, divide the interval into 2 ^{n} equal pieces and deﬁne the step functions to take inﬁmim of the continuous function on the corresponding interval. Then use uniform convergence.
(2) By using the ‘telescoping trick’ show that any continuous function on [0, 1) can be written as the sum
(3.17)
� f _{j} (x) ∀ x ∈ [0, 1)
i
where the f _{j} are step functions and
�
j
f _{j} (x) < ∞ for all x ∈ [0, 1).
(3) Conclude that any continuous function on [0, 1], extended to be 0 outside this interval, is a Lebesgue integrable function on R.
14
LECTURE NOTES FOR 18.102, SPRING 2009
Solutions to Problem set 1
Full marks will be given to anyone who makes a good faith attempt to answer each question. The ﬁrst four problems concern the ‘little L p’ spaces l ^{p} . Note that you have the choice of doing everything for p = 2 or for all 1 ≤ p < ∞. Everyone who handed in a script received full marks. Problem 1.1 Write out a proof (you can steal it from one of many places but at least write it out in your own hand) either for p = 2 or for each p with 1 ≤ p < ∞ that
l ^{p} = {a : N −→ C;
_{∞}
�
j=1
a _{j}  ^{p} < ∞,
is a normed space with the norm
⎛
�a� _{p} = ⎝
∞
�
j=1
a _{j} 
p
⎞
⎠
1
p
.
a _{j} = a(j)}
This means writing out the proof that this is a linear space and that the three conditions required of a norm hold. Solution: We know that the functions from any set with values in a linear space form a linear space – under addition of values (don’t feel bad if you wrote this out, it is a good thing to do once). So, to see that l ^{p} is a linear space it suﬃces to see that it is closed under addition and scalar multiplication. For scalar multiples this is clear:
ta _{i}  = ta _{i}  so �ta� _{p} = t�a� _{p}
(3.18)
which is part of what is needed for the proof that � · � _{p} is a norm anyway. The fact that a, b ∈ l ^{p} imples a + b ∈ l ^{p} follows once we show the triangle inequality or we can be a little cruder and observe that
(3.19)
a _{i} + b _{i}  ^{p} ≤ (2
max(a _{i} , b _{i} )) ^{p} = 2 ^{p} max(a ^{p} , b _{i}  ^{p} ) ≤ 2 ^{p} (a _{i}  + b _{i} )
i
_{�}_{a} _{+} _{b}_{�} _{p} ^{p} = ^{�} a _{i} + b _{i}  ^{p} ≤ 2 ^{p} (�a� ^{p} + �b� ^{p} ),
j
where we use the fact that t ^{p} is an increasing function of t ≥ 0. Now, to see that l ^{p} is a normed space we need to check that �a� _{p} is indeed a norm. It is nonnegative and �a� _{p} = 0 implies a _{i} = 0 for all i which is to say a = 0. So, only the triangle inequality remains. For p = 1 this is a direct consequence of the usual triangle inequality:
�
(3.20)
�a + b� _{1} =
�
a _{i} + b _{i}  ≤
(a _{i}  + b _{i} ) = �a� _{1} + �b� _{1} .
i
i
For 1 < p < ∞ it is known as Minkowski’s inequality. This in turn is deduced from H¨older’s inequality – which follows from Young’s inequality! The latter says if 1/p + 1/q = 1, so q = p/(p − 1), then
(3.21)
αβ ≤
α
p
p
+
β
q
q
∀ α, β ≥ 0.
To check it, observe that as a function of α = x,
(3.22)
f (x) =
x
p
p
− xβ +
β ^{q}
q
LECTURE NOTES FOR 18.102, SPRING 2009
15
if nonnegative at x = 0 and clearly positive when x >> 0, since x ^{p} grows faster than xβ. Moreover, it is diﬀerentiable and the derivative only vanishes at x ^{p}^{−}^{1} = β, where it must have a global minimum in x > 0. At this point f (x) = 0 so
Young’s inequality follows. Now, applying this with α = a _{i} /�a� _{p} and β = b _{i} /�b� _{q} (assuming both are nonzero) and summing over i gives H¨older’s inequality
(3.23)
�

i
a _{i} b _{i} /�a� _{p} �b� _{q} ≤
�
i
=⇒ 
a _{i} b _{i} /�a� _{p} �b� _{q} ≤
�
i
� a _{i} b _{i}  ≤ �a� _{p} �b� _{q} .
i
�
b p ^{+} � b � ^{q} q
a
i
p
�
a �
p
p
i
q
q




�
= 1
Of course, if either �a� _{p} = 0 or �b� _{q} = 0 this inequality holds anyway. Now, from this Minkowski’s inequality follows. Namely from the ordinary trian gle inequality and then Minkowski’s inequality (with q power in the ﬁrst factor)
(3.24)
�
i
a _{i} + b _{i}  ^{p} =
�
i
a _{i} + b _{i}  ^{(}^{p}^{−}^{1}^{)} a _{i} + b _{i} 
�
≤
i
a _{i} + b _{i}  ^{(}^{p}^{−}^{1}^{)} a _{i}  +
�
i
≤
a _{i} + b _{i}  ^{(}^{p}^{−}^{1}^{)} b _{i} 
�
�
i
a _{i} + b _{i}  ^{p}
^{�}
1/q
(�a� _{p} + �b� _{q} )
gives after division by the ﬁrst factor on the right
(3.25)
Thus, l ^{p} is indeed a normed space. I did not necessarily expect you to go through the proof of YoungH¨older Minkowksi, but I think you should do so at some point since I will not do it in class. Problem 1.2 The ‘tricky’ part in Problem 1.1 is the triangle inequality. Suppose you knew – meaning I tell you – that for each N
�a + b� _{p} ≤ �a� _{p} + �b� _{p} .
⎛
⎝
N
�
j=1
a _{j} 
⎞ p ⎠
1
p
is a norm on C ^{N}
would that help? Solution: Yes indeed it helps. If we know that for each N
(3.26)
⎛
⎝
N
�
j=1
a _{j} + b _{j} 
p
⎞
⎠
1
^{p}
≤
⎛
⎝
N
�
j=1
a _{j}  ^{p}
⎞
⎠
1
^{p}
+
⎛
⎝
N
�
j=1
b _{j} 
p
⎞
⎠
1
^{p}
then for elements of l ^{p} the norms always bounds the right side from above, meaning
(3.27)
⎛
⎝
N
�
j=1
a _{j} + b _{j} 
p
⎞
⎠
1
p
≤
�a� _{p} + �b� _{p} .
Since the left side is increasing with N it must converge and be bounded by the right, which is independent of N. That is, the triangle inequality follows. Really
16
LECTURE NOTES FOR 18.102, SPRING 2009
this just means it is enough to go through the discussion in the ﬁrst problem for ﬁnite, but arbitrary, N. Problem 1.3 Prove directly that each l ^{p} as deﬁned in Problem 1.1 – or just l ^{2} – is complete, i.e. it is a Banach space. At the risk of oﬀending some, let me say that this means showing that each Cauchy sequence converges. The problem here is to ﬁnd the limit of a given Cauchy sequence. Show that for each N the sequence in C ^{N} obtained by truncating each of the elements at point N is Cauchy with respect to the norm in Problem 1.2 on C ^{N} . Show that this is the same as being Cauchy in C ^{N} in the usual sense (if you are doing p = 2 it is already the usual sense) and hence, this cutoﬀ sequence converges. Use this to ﬁnd a putative limit of the Cauchy sequence and then check that it works. Solution: So, suppose we are given a Cauchy sequence a ^{(}^{n}^{)} in l ^{p} . Thus, each
element is a sequence {a
in l ^{p} . From the continuity of the norm in Problem
1.5 below, �a ^{(}^{n}^{)} � must be Cauchy in R and so converges. In particular the sequence is norm bounded, there exists A such that �a ^{(}^{n}^{)} � _{p} ≤ A for all n. The Cauchy condition itself is that given � > 0 there exists M such that for all m, n > M,
(n) _{}}
j
∞
j=1
(3.28)
�a ^{(}^{n}^{)} − a ^{(}^{m}^{)} � _{p} =
�
�
i
a
(n)
i
− a
(m)
i

p
�
1
p
< �/2.
Now for each i, a _{i}
(n) _{−} _{a} i
^{(}^{m}^{)}  ≤ �a ^{(}^{n}^{)} − a ^{(}^{m}^{)} � _{p} so each of the sequences a ^{(}^{n}^{)} _{i} must
be Cauchy in C. Since C is complete
(3.29)
So, our putative limit is a, the sequence {a _{i} } shows that
lim a ^{(}^{n}^{)}
_{i}
n→∞
= a _{i} exists for each i = 1, 2,
∞
i=1 ^{.}
The boundedness of the norms
(3.30)
N
�
i=1
^{(}^{n}^{)}  ^{p} ≤ A ^{p}
a _{i}
and we can pass to the limit here as n → ∞ since there are only ﬁnitely many terms. Thus
(3.31)
N
�
i=1
a _{i}  ^{p} ≤ A ^{p}
∀
N =⇒ �a� _{p} ≤ A.
Thus, a ∈ l ^{p} as we hoped.
ﬁnite inequality which follows from the Cauchy conditions
Similarly, we can pass to the limit as m → ∞ in the
(3.32)
to see that for each N
(3.33)
(
N
�
i=1
a
(n)
i
− a
(m)
i

(
N
�
i=1
a
(n)
i
− a
i 
p
p
)
)
1
p
1
p
< �/2
≤ �/2
and hence
(3.34)
Thus indeed, a ^{(}^{n}^{)} →
�a ^{(}^{n}^{)} − a� < � ∀ n > M.
a in l
p as we were trying to show.
LECTURE NOTES FOR 18.102, SPRING 2009
17
Notice that the trick is to ‘back oﬀ’ to ﬁnite sums to avoid any issues of inter changing limits. Problem 1.4 Consider the ‘unit sphere’ in l ^{p} – where if you want you can set p = 2. This is the set of vectors of length 1 :
S = {a ∈ l ^{p} ;�a� _{p} = 1}.
(1) Show that S is closed. (2) Recall the sequential (so not the open covering deﬁnition) characterization of compactness of a set in a metric space (e.g. by checking in Rudin). (3) Show that S is not compact by considering the sequence in l ^{p} with kth element the sequence which is all zeros except for a 1 in the kth slot. Note that the main problem is not to get yourself confused about sequences of sequences! Solution: By the next problem, the norm is continuous as a function, so
(3.35)
is the inverse image of the closed subset {1}, hence closed. Now, the standard result on metric spaces is that a subset is compact if and only if every sequence with values in the subset has a convergent subsequence with limit in the subset (if you drop the last condition then the closure is compact). In this case we consider the sequence (of sequences)
S = {a; �a� = 1}
(3.36)
(n)
a _{i}
=
�
0
1
�
i = n
.
i = n
1
This has the property that �a ^{(}^{n}^{)} − a ^{(}^{m}^{)} � _{p} = 2 p whenever n =� m. Thus, it cannot have any Cauchy subsequence, and hence cannot have a convergent subsequence, so S is not compact. This is important. In fact it is a major diﬀerence between ﬁnitedimensional and inﬁnitedimensional normed spaces. In the latter case the unit sphere cannot be compact whereas in the former it is. Problem 1.5 Show that the norm on any normed space is continuous. Solution: Right, so I should have put this problem earlier! The triangle inequality shows that for any u, v in a normed space
(3.37) 
�u� ≤ 
�u − v� + �v�, 
�v� ≤ �u − v� + �u� 
which implies that 

(3.38) 
�u� − �v� ≤ �u − v�. 
This shows that � · � is continuous, indeed it is Lipschitz continuous.
MIT OpenCourseWare http://ocw.mit.edu
18.102 Introduction to Functional Analysis
Spring 2009
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
18
LECTURE NOTES FOR 18.102, SPRING 2009
Lecture 4. Thursday, 12 Feb
I talked about step functions, then the covering lemmas which are the basis of the deﬁnition of Lebesgue measure – which we will do after the integral – then properties of monotone sequences of step functions. To be deﬁnite, but brief, by an interval we will mean [a, b) – an interval closed on the left and open on the right – at least for a little while. This is just so the length of the interval, b − a, only vanishes when the interval is empty (not true for closed intervals of course) and also so that we can decompose an interval, in this sense, into two disjoint intervals by choosing any interior point:
(4.1)
Now, by a step function
(4.2)
f : R −→ C
[a, b) = [a, t) ∪ [t, b), a < t < b.
(although often we will restrict to functions with real values) we mean a function which vanishes outside a ﬁnite union of disjoint ‘intervals’ and is constant on each of them. Thus f (R) is ﬁnite – the function only takes ﬁnitely many values – and
(4.3)
f ^{−}^{1} (c) is a ﬁnite union of disjoint intervals, c �= 0.
It is also often convenient to write a step function as a sum
(4.4)
f =
N
�
i=1
^{c} i ^{χ} [a _{i} ,b _{i} )
of multiples of the characteristic functions of our intervals. Note that such a ‘pre sentation’ is not unique but can be made so by demanding that the intervals be disjoint and ‘maximal’ – so f is does not take the same value on two intervals with a common endpoint. Now, a constant multiple of a step function is a step function and so is the sum
of two step functions – clearly the range is ﬁnite. Really this reduces to checking that the diﬀerence [a, b) \ [a ^{�} , b ^{�} ) and the union of two intervals is always a union of intervals. The absolute value of a step function is also a step function. A similar argument shows that the integral, deﬁned by
(4.5)
�
R
f =
�
i
c _{i} (b _{i} − a _{i} )
from (4.4) is independent of the ‘presentation’ used to deﬁne it. It is of course equal to the Riemann integral which is one way of seeing that it is welldeﬁned (but of course the result is much more elementary).
Proposition 1. The step functions on R in the sense deﬁned above form a normed space with the L ^{1} norm
(4.6)
�f � _{L} 1 =
�
R
f.
So, we will complete this space instead of the continuous functions – it is both more standard and a little easier. The fact that we can directly construct a ‘con crete’ completion is due to Mikusinski.´ Already at this stage we can deﬁne a Lebesgue integrable function, however we need to do some work to ﬂesh out the deﬁnition.
LECTURE NOTES FOR 18.102, SPRING 2009
19
Deﬁnition 3. A function g : R −→ C is Lebesgue integrable if there exists an absolutely summable sequence of step functions f _{n} , i.e. satisfying
(4.7) 

such that 

(4.8) 
f (x) = 
�
n
�
n
� f _{n}  < ∞
f _{n} (x) ∀ x ∈ R such that
�
n
f _{n} (x) < ∞.
So, the deﬁnition is a little convoluted – there must exist a sequence of step functions the sum the integrals of the absolute values of which converges and such that the sum itself converges to the function but only at the points of absolute converge of the (pointwise) series. Tricky, but one can get used to it – and ultimately simplify it. The main attraction of this deﬁnition is that it is selfcontained and in principle ‘everything’ can be deduced from it. Now, the ﬁrst thing we need is the covering lemma – basically some properties of countable collections of our ‘intervals’ such as will arise when we look at a sequence, or series, of step functions. I leave it to you to devise careful proofs of the following two facts.
Lemma 1. If C _{i} = [a _{i} , b _{i} ), i = 1,
, N, is a
ﬁnite collection of intervals then
(4.9)
C _{i} ⊂ [a, b) ∀ i and C _{i} ∩ C _{j} = ∅ ∀ i =�
j =⇒
N
�
i=1
(b _{i} − a _{i} ) ≤ (b − a).
On the other hand
(4.10)
[a, b) ⊂
N
C _{i} =⇒
i=1
N
�
i=1
(b _{i} − a _{i} ) ≥ (b − a).
You can prove this by inserting division points etc. Now, what we want is the same thing for a countable collection of intervals.
Proposition 2. If C _{i} = [a _{i} , b _{i} ), i ∈ N, is a countable collection of intervals then
(4.11)
or
(4.12)
C _{i} ⊂ [a, b) ∀ i and C _{i} ∩ C _{j} = ∅ ∀ i =�
j =⇒
∞
�
i=1
(b _{i} − a _{i} ) ≤ (b − a)
[a, b) ⊂
N
C _{i} =⇒
i=1
∞
� (b _{i} − a _{i} ) ≥ (b − a).
i=1
Proof. You might think these are completely obvious, and the ﬁrst is – the hypoth esis (4.9) holds for any ﬁnite subcollection and hence the ﬁnite sum is always less than the ﬁxed number b − a and hence so is the inﬁnite sum – which therefore converges. On the other hand (4.12) is not quite so obvious since it depends on HeineBorel. To be able to apply (4.10) choose δ > 0. Now extend each interval by replacing the lower limit by a _{i} − 2 ^{−}^{i} δ and consider the open intervals which therefore have the property
(4.13)
[a, b − δ] ⊂ [a, b) ⊂
(a _{i} − 2 ^{−}^{i} δ, b _{i} ).
i
20
LECTURE NOTES FOR 18.102, SPRING 2009
Now, by HeineBorel – the compactness of closed bounded intervals – a ﬁnite subcol lection of these open intervals covers [a, b − δ) so (4.10) does apply to the semiopen intervals and shows that for some ﬁnite N (hence including the ﬁnite subcollection)
(4.14)
N
�
i=1
(b _{i} − a _{i} ) − 2 ^{−}^{i} δ ≥ b − a − δ =⇒
∞
�
i=1
(b _{i} − a _{i} ) ≥ b − a − 2δ
where the sum here might be inﬁnite, but then it is all the more true. The fact
�
that this is true for all δ > 0 now proves (4.12).
Welcome to measure theory. Okay, now the basis result on which most of the properties of integrable functions hinge is the following monotonicity lemma.
Lemma 2. Let f _{n} be a sequence of step functions which decreases monotonically to 0
(4.15)
then
f _{n} (x) ↓ 0 ∀ x ∈ R,
(4.16)
lim
�
f _{n} = 0.
Note that ‘decreases’ here means that f _{n} (x) is, for each x, a nonincreasing sequence which has limit 0. Of course this means that all the f _{n} are nonnegative. Moreover the ﬁrst one vanishes outside some interval [a, b) hence so do all of them. The crucial thing about this lemma is that we get the vanishing of the limit of the sequence of integrals without having to assume uniformity of the limit.
Proof. Going back to the deﬁnition of the integral of step functions, clearly f _{n} (x) ≥
�
f _{n}_{+}_{1} (x) for all x implies that f _{n} is a decreasing (meaning nonincreasing) se
quence. So there are only two possibilities, it converges to 0, as we claim, or it
converges to some positive value. This means that there is some δ > such that
� f _{n} > δ for all n, so we just need to show that this is not so. Given an � > 0 consider the sets
(4.17)
S _{j} = {x ∈ [a, b); f _{j} (x) ≤ �}.
(Here [a, b) is an interval outside which f _{1} vanishes and hence all the f _{n} vanish outside it). Each of the S _{j} is a ﬁnite union of intervals. Moreover
(4.18)
S _{j} = [a, b)
j
since f _{n} (x) and
(4.19)
then the B _{j} are all disjoint, each consists of ﬁnitely many intervals and
(4.20)
Thus, both halves of Proposition 2 apply to the intervals forming the B _{j} . If we let l(B _{j} ) b
Гораздо больше, чем просто документы.
Откройте для себя все, что может предложить Scribd, включая книги и аудиокниги от крупных издательств.
Отменить можно в любой момент.