Math 209B Homework 2

Similar documents
THEOREMS, ETC., FOR MATH 515

Continuous Functions on Metric Spaces

Real Analysis, 2nd Edition, G.B.Folland Elements of Functional Analysis

7 Complete metric spaces and function spaces

Problem Set 5: Solutions Math 201A: Fall 2016

Real Analysis Chapter 4 Solutions Jonathan Conder

Analysis III Theorems, Propositions & Lemmas... Oh My!

Problem Set 5. 2 n k. Then a nk (x) = 1+( 1)k

Folland: Real Analysis, Chapter 4 Sébastien Picard

ANALYSIS WORKSHEET II: METRIC SPACES

Analysis III. Exam 1

Overview of normed linear spaces

Functional Analysis Exercise Class

Real Analysis Notes. Thomas Goller

Lecture Notes in Advanced Calculus 1 (80315) Raz Kupferman Institute of Mathematics The Hebrew University

Math General Topology Fall 2012 Homework 11 Solutions

1 Topology Definition of a topology Basis (Base) of a topology The subspace topology & the product topology on X Y 3

Notions such as convergent sequence and Cauchy sequence make sense for any metric space. Convergent Sequences are Cauchy

Compact operators on Banach spaces

Math 328 Course Notes

converges as well if x < 1. 1 x n x n 1 1 = 2 a nx n

Notes for Functional Analysis

REAL AND COMPLEX ANALYSIS

THE STONE-WEIERSTRASS THEOREM AND ITS APPLICATIONS TO L 2 SPACES

CONTENTS. 4 Hausdorff Measure Introduction The Cantor Set Rectifiable Curves Cantor Set-Like Objects...

1. Let A R be a nonempty set that is bounded from above, and let a be the least upper bound of A. Show that there exists a sequence {a n } n N

(1) Consider the space S consisting of all continuous real-valued functions on the closed interval [0, 1]. For f, g S, define

Metric Spaces and Topology

Real Analysis Math 131AH Rudin, Chapter #1. Dominique Abdi

L p Spaces and Convexity

MAA6617 COURSE NOTES SPRING 2014

INTRODUCTION TO TOPOLOGY, MATH 141, PRACTICE PROBLEMS

MA677 Assignment #3 Morgan Schreffler Due 09/19/12 Exercise 1 Using Hölder s inequality, prove Minkowski s inequality for f, g L p (R d ), p 1:

McGill University Math 354: Honors Analysis 3

THEOREMS, ETC., FOR MATH 516

MAT 570 REAL ANALYSIS LECTURE NOTES. Contents. 1. Sets Functions Countability Axiom of choice Equivalence relations 9

MH 7500 THEOREMS. (iii) A = A; (iv) A B = A B. Theorem 5. If {A α : α Λ} is any collection of subsets of a space X, then

Notes for Functional Analysis

Lectures on Analysis John Roe

Immerse Metric Space Homework

Most Continuous Functions are Nowhere Differentiable

Functional Analysis. Franck Sueur Metric spaces Definitions Completeness Compactness Separability...

Chapter 1. Introduction

Locally convex spaces, the hyperplane separation theorem, and the Krein-Milman theorem

FUNCTIONAL ANALYSIS-NORMED SPACE

Continuity. Matt Rosenzweig

Exercise Solutions to Functional Analysis

The Arzelà-Ascoli Theorem

Summer Jump-Start Program for Analysis, 2012 Song-Ying Li. 1 Lecture 7: Equicontinuity and Series of functions

MTH 503: Functional Analysis

Lecture 5 - Hausdorff and Gromov-Hausdorff Distance

Your first day at work MATH 806 (Fall 2015)

Math 426 Homework 4 Due 3 November 2017

REVIEW OF ESSENTIAL MATH 346 TOPICS

2 (Bonus). Let A X consist of points (x, y) such that either x or y is a rational number. Is A measurable? What is its Lebesgue measure?

3 (Due ). Let A X consist of points (x, y) such that either x or y is a rational number. Is A measurable? What is its Lebesgue measure?

Introduction to Topology

d(x n, x) d(x n, x nk ) + d(x nk, x) where we chose any fixed k > N

Metric Spaces Math 413 Honors Project

E.7 Alaoglu s Theorem

Spaces of continuous functions

PROBLEMS. (b) (Polarization Identity) Show that in any inner product space

Real Analysis Problems

Problem Set 2: Solutions Math 201A: Fall 2016

Analysis Qualifying Exam

Continuity of convex functions in normed spaces

Principles of Real Analysis I Fall VII. Sequences of Functions

Math 5052 Measure Theory and Functional Analysis II Homework Assignment 7

16 1 Basic Facts from Functional Analysis and Banach Lattices

Topology. Xiaolong Han. Department of Mathematics, California State University, Northridge, CA 91330, USA address:

Math 118B Solutions. Charles Martin. March 6, d i (x i, y i ) + d i (y i, z i ) = d(x, y) + d(y, z). i=1

MATH 51H Section 4. October 16, Recall what it means for a function between metric spaces to be continuous:

CHAPTER I THE RIESZ REPRESENTATION THEOREM

l(y j ) = 0 for all y j (1)

The Heine-Borel and Arzela-Ascoli Theorems

A VERY BRIEF REVIEW OF MEASURE THEORY

1/12/05: sec 3.1 and my article: How good is the Lebesgue measure?, Math. Intelligencer 11(2) (1989),

Introduction and Preliminaries

Linear Analysis Lecture 5

Bounded and continuous functions on a locally compact Hausdorff space and dual spaces

Contents Ordered Fields... 2 Ordered sets and fields... 2 Construction of the Reals 1: Dedekind Cuts... 2 Metric Spaces... 3

An introduction to some aspects of functional analysis

MTG 5316/4302 FALL 2018 REVIEW FINAL

Math 421, Homework #9 Solutions

MAT 578 FUNCTIONAL ANALYSIS EXERCISES

Bootcamp. Christoph Thiele. Summer As in the case of separability we have the following two observations: Lemma 1 Finite sets are compact.

Functional Analysis HW #1

Functional Analysis Winter 2018/2019

4 Countability axioms

Economics 204 Summer/Fall 2011 Lecture 5 Friday July 29, 2011

Assignment-10. (Due 11/21) Solution: Any continuous function on a compact set is uniformly continuous.

THE INVERSE FUNCTION THEOREM

SOME QUESTIONS FOR MATH 766, SPRING Question 1. Let C([0, 1]) be the set of all continuous functions on [0, 1] endowed with the norm

Functional Analysis I

David Hilbert was old and partly deaf in the nineteen thirties. Yet being a diligent

Chapter 3: Baire category and open mapping theorems

Chapter 1: Banach Spaces

Analysis Comprehensive Exam Questions Fall F(x) = 1 x. f(t)dt. t 1 2. tf 2 (t)dt. and g(t, x) = 2 t. 2 t

Review Notes for the Basic Qualifying Exam

CHAPTER V DUAL SPACES

Transcription:

Math 29B Homework 2 Edward Burkard Note: All vector spaces are over the field F = R or C 4.6. Two Compactness Theorems. 4. Point Set Topology Exercise 6 The product of countably many sequentally compact spaces is sequentially compact. (Use the diagonal trick as in the proof of Theorem 4.44.) Exercise 64 Let (X, ρ) be a metric space. A function f C(X) is called Hölder continuous of exponent α (α > ) if the quantity f(x) f(y) N α (f) = sup x y ρ(x, y) α is finite. If X is compact, {f C(X) f u and N α (f) } is compact in C(X). Let F = {f C(X) f u and N α (f) }. Let f F. Then we have that N α (f) implies that f(x) f(y) ρ(x, y) α. Let ɛ > and let x X. Then to show equicontinuity, let U = B (x), hence: ɛ α f(x) f(y) (ɛ α ) α = ɛ which implies, since x was arbitrary, F is equicontinuous on X. Recall that f u = sup x X f(x). So since every function in F satisfies f u it is clear that F is uniformly bounded. So by the Arzelà-Ascoli Theorem I we have that, since X is compact and Hausdorff (since it is a metric space), the closure of F in C(X) is compact. So to prove the claim it remains to show that F is closed. Let {f n } be a sequence in F converging to some function f. Observe: f u = lim f n u = sup lim f n (x) x X = sup lim f n (x) x X = lim sup f n (x) x X = lim = N α (f) = sup x y lim f n (x) lim f n (y) ρ α (x, y) lim(f n (x) f n (y)) = sup x y ρ α (x, y) = sup lim (f n(x) f n (y)) x y ρ α (x, y) (f n (x) f n (y)) = lim sup x y ρ α (x, y) = lim = Hence F is closed, and therefore F is compact in C(X).

2 4.7. The Stone-Weierstrass Theorem. Exercise 68 Let X and Y be compact Hausdorff spaces. The algebra generated by functions of the form f(x, y) = g(x)h(y), where g C(X) and h C(Y ), is dense in C(X Y ). Since both X and Y are compact and Hausdorff, by Tychonoff s theorem and Proposition 4. we have that X Y is compact and Hausdorff. Now consider the algebra described above. Call it A. Clearly A contains the constant functions as we can let g and h both be constants and hence we have an f(x, y) = r for all r R. Also clearly A separates points (this can be done by letting g(x) =, h(x) = 2, k(y) = 3 and given any two points (a, b) and (c, d) plugging them into f (x, y) = g(x)k(y) = 3 and f 2 (x, y) = h(x)k(y) = 6 respectively). Since A is closed by definition, we can apply the Stone-Weierstrass Theorem to it to conclude that A is dense in C(X Y ). 5.. Normed Vector Spaces. Exercise 3 Complete the proof of Proposition 5.4. 5. Elements of Functional Analysis We need to verify that T := lim T n is linear and T n T. First: Hence T is linear. Now check the norm of T : T (ax + by) = lim T n (ax + by) = lim(t n (ax) + T n (by)) = lim(at n (x) + bt n (y)) = lim(at n (x)) + lim(bt n (y)) = a lim T n (x) + b lim T n (y) = at (x) + bt (y) T = sup{ T x x = } = sup{ lim n T nx x = } = sup{ lim n T nx x = } = lim n sup{ T nx x = } = lim n T n Now fix x X and let ɛ >, then since {T n } is Cauchy there is an N N such that for all m, n N: Notice: T n (x) T m (x) ɛ T n (x) T (x) = T n (x) lim m T m(x) = lim m (T n(x) T m (x)) = lim m T n(x) T m (x) < ɛ Letting ɛ we see that T n (x) T (x). So since x was arbitrary we have T n T. Exercise 6 Suppose that X is a finite dimensional vector space. Let e,..., e n be a basis for X, and define n a je j = n a j. is a norm on X. The map (a,..., a n ) n a je j is continuous from F n with the usual Euclidean topology to X with the topology defined by. c. {x X x = } is compact in the topology defined by. d. All norms on X are equivalent. (Compare any norm to.)

3 a j e j + b j e j = = n λ a j e j = a j + b j a j + b j a j e j + b j e j λa j = λ a j = λ a j e j Now if n a je j =, then we have n a j =, which is a sum of nonnegative numbers, hence a j = for all j, thus n a je j =. So we conclude that is a norm. Define the map η : F n X by η(a,..., a n ) = a j e j. Using the norm (a,..., a n ) = a j on F n suggested by Folland on page 53 we can see: (a,..., a n ) = a j = η(a,..., a n ) = a j e j which makes it obvious that η is continuous. c. Let K = {(a i,..., a j ) (a,..., a j ) = }. Then K is clearly compact in F n. Now since we have η(k) = {x X x = } (by the definition of η) we can see that η(k) is compact as it is the continuous image of a compact set. Thus we have our conclusion. d. Let be any norm on X. Define M = max( e,..., e n ). Then we have a j e j a j e j = a j e j M a j = M a j e j. Thus we have x M x for all x X. As this is true, we can conclude that the map sending one norm to the other, in this case use ζ : (X, ) (X, ), is continuous (this is really just a dialation of the space). Recall the set K from the previous part. Since K is compact and ζ is continuous, ζ(k) is compact. By the extreme value theorem ζ assumes a minimum value as measured by, so we may as well say assumes a minimum value on K, call it m. Since is a norm, m and since x / K, m >. Given any element x X, we can construct the element x which is in K and hence x x m which implies m x x and so: m x x M x. Since this holds for any x X we have that is equivalent to. Moreover, since was arbitrary, every norm on X is equivalent to. Exercise 7 Let X be a Banach space. If T L(X, X ) and I T < where I is the identity operator, then T is invertible; in fact, the series (I T )n converges in L(X, X ) to T. If T L(X, X ) is invertible and S T < T, then S is invertible. Thus the set of invertible operators is open in L(X, X ).

4 Since X is a Banach space we have that it is complete, and hence by Proposition 5.4 L(X, X ) is complete. Let T L(X, X ) be such that I T <. Consider the sum (I T )n which is in L(X, X ) since it is an absolutely convergent sum (since I T < and hence I T n is a convergent geometric series) by Theorem 5.. Notice that showing I = T (I T )n proves that T exists and in fact T = (I T )n. A simple use of induction shows that T N (I T )n = I (I T ) N+. Now notice: N I (T (I T ) n ) = I (I (I T )N+ ) = (I T ) N+ I T N + which gives that I (T N (I T )n ) as N since I T <, hence I T (I T ) n = and so or equivalently T (I T ) n = I T = (I T ) n. Thus T is invertible. Let T L(X, X ) be invertible and let S L(X, X ) be such that S T < T (an equivalent statement: T S T < ). So: I ST = T (T S) T S T <. By the proof in part a, we have that ST is invertible, hence: I = (ST )(ST ) = S(T (ST ) ) and thus S = T (ST ). This proves that for any S in L(X, X ) sufficiently close to T, S is invertible, hence the set of invertible operators is open in L(X, X ). Exercise If < α, let Λ α ([, ]) be the space of Hölder continuous functions of exponent α on [, ]. That is, f Λ α ([, ]) iff f Λα <, where f(x) f(y) f Λα = f() + sup x,y [,], x y x y α. Λα is a norm that makes Λ α ([, ]) into a Banach space. Let λ α ([, ]) be the set of all f Λ α ([, ]) such that f(x) f(y) x y α as x y, for all y [, ]. If α <, λ α ([, ]) is an infinite-dimensional closed subspace of Λ α ([, ]). If α =, λ α ([, ]) contains only constant functions. Exercise 2 Let X be a normed vector space and M a proper closed subspace of X. x + M = inf{ x + y y M} is a norm on X /M. For any ɛ > there exists x X such that x = and x + M ɛ. c. The projection map π(x) = x + M from X to X /M has norm. d. If X is complete, so is X /M. (Use Theorem 5.) e. The topology defined by the quotient norm is the quotient topology as defined in Exercise 28.

5 Let x + M, y + M X /M. Then: (x + y) + M = inf{ (x + y) + z z M} { = inf (x + 2 ) ( z + y + ) } 2 z z M {( since is a norm on X inf x + ) ( 2 z + y + ) } 2 z z M let w = z 2 = inf{ x + w w M} + inf{ y + z w M} = x + M + y + M Thus since x and y were arbitrary, satisfies the triangle inequality for all x + M, y + M X /M. Now let x + M X /M and let c R. Then: c(x + M) = inf{ c(x + y) y M} since is a norm on X = inf{ c x + y y M} = c inf{ x + y y M} = c x + M Now suppose that x + M =. Then clearly x M, and hence x + M = in X /M, because if it were not, then it would be impossible for x + M to be as we are taking the infimum over vectors in M and x / M. Thus it is a norm on X /M. As M is a proper closed subset, we can find an x X /M such that x + M = a >. Define y = ax, then y + M =. Now by the definition of infimum we have that there is some m M with y + m + ɛ 2. Thus there is a element Υ B ɛ(y + m) with Υ =. So we have y + M Υ + M Υ y + M = Υ (y + m) + M Υ (y + m) ɛ Yielding the inequality: or equivalently c. Observe: d. e. y + M Υ + M = Υ + M ɛ π(x) = x + M = ɛ Υ + M. { x if x / M if x M Hence π(x) x which means π. However part b tells us π and hence π =. 5.2. Linear Functionals. Exercise 7 A linear functional f on a normed vector space X is bounded iff f ({}) is closed. (Use Exercise 2) (= ) Assume that f is bounded. Then it is continuous. So since {} is closed in R and f is continuous, f ({}) is closed as desired. ( =) Suppose that f were not bounded. Then n > there is an x n X with x n = and f(x n ) > 2 n. ( ) Define y n = xn f(x. Then y x n) n = n f(x = n) f(x x n) 2 and f(y n n ) = f n f(x n) Now fix x X and consider the sequence a n = x f(x)y n. Then a n x as n since y n as n. However, f(a n ) = f(x) f(x)f(y n ) = f(x) f(x) =. Thus we have a n f ({}). This would imply that f ({}) is dense in X. So there are two choices, either f : X {} or f ({}) is not closed. Since the first choice is obviously not true, it must be that f ({}) is not closed. Hence by contrapositive we have our desired conclusion. = f(x n) f(x n) =.

6 Exercise 8 Let X be a normed vector space. If M is a closed subspace and x X /M then M + Cx is closed. (Use Theorem 5.8) Every finite-dimensional subspace of X is closed. Exercise 9 Let X be an infinite-dimensional normed vector space. There is a sequence {x j } in X such that x j = for all j and x j x k 2 x j inductively, using Exercised 2b and 8.) X is not locally compact. for j k. (Construct Exercise 25 If X is a Banach space and X is separable, then X is separable. (Let {f n } be a countable dense subset of X. For each n choose x n X with x n = and f n (x n ) 2 f n. Then the linear combinations of {x n } are dense in X.) Note: Separability of X does not imply separability of X. 5.3. The Baire Category Theorem and Its Consequences. Exercise 27 There exist meager subsets of R whose complements have Lebesgue measure zero. Exercise 29 Let Y = L (µ) where µ is counting measure on N, and let X = {f SCY {n f(n) < }, equipped with the L norm. X is a proper dense subspace of Y ; hence X is not complete. Define T : X Y by T f(n) = nf(n). Then T is closed but not bounded. c. Let S = T. Then S : Y X is bounded and surjective, but not open. c. Exercise 3 Let Y = C([, ]) and X = C ([, ]), both equipped with the uniform norm. X is not complete. The map ( d dx) : X Y is closed (see Exercise 9) but not bounded. Consider the function f(x) = x 2 defined on [, ]. By lemma 4.47, ɛ > there is a polynomial P (x) defined on R such that P () = and x P (x) u < ɛ for x [, ]. If we let Q(x) = P ( ) x 2 we can rephrase the lemma to say ɛ > there is a polynomial Q(x) such that Q ( 2) = and x 2 Q(x) u < ɛ for x [, ]. Since these Q s are polynomials they are clearly in C ([, ]), and thus we may create a sequence of them as follows: Define Q n (x) to be one of the polynomials such that x 2 Q n (x) u < n for all n N. This creates a convergent sequence {Q n } in C ([, ]) which converges to x 2 on [, ], however x 2 / C ([, ]). Thus C ([, ]) is not complete. Exercise 32 Let and 2 be norms on the vector space X such that 2. If X is complete with respect to both norms, then the norms are equivalent.

5.4. Topological Vector Spaces. Exercise 45 The space C (R) of all infinitely differentiable functions on R has a Fréchet space topology with respect to which f n f iff f (k) n f (k) uniformly on compact sets for all k. Exercise 47 Suppose that X and Y are Banach spaces. If {T n } L(X, Y ) and T n T weakly (or strongly), then sup N T n <. Every weakly convergent sequence in X, and every weak -convergent sequence in X, is bounded (with respect to the norm). 7