SOLUTIONS OF SELECTED PROBLEMS

Similar documents
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?

Advanced Analysis Qualifying Examination Department of Mathematics and Statistics University of Massachusetts. Tuesday, January 16th, 2018

THEOREMS, ETC., FOR MATH 515

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?

A List of Problems in Real Analysis

Measurable functions are approximately nice, even if look terrible.

Chapter 5. Measurable Functions

MATHS 730 FC Lecture Notes March 5, Introduction

Math 5051 Measure Theory and Functional Analysis I Homework Assignment 3

MAT 571 REAL ANALYSIS II LECTURE NOTES. Contents. 2. Product measures Iterated integrals Complete products Differentiation 17

Integration on Measure Spaces

FINAL REVIEW Answers and hints Math 311 Fall 2017

Math 140A - Fall Final Exam

Lebesgue Integration on R n

MATH 202B - Problem Set 5

Real Analysis Notes. Thomas Goller

02. Measure and integral. 1. Borel-measurable functions and pointwise limits

CHAPTER 6. Differentiation

Homework 11. Solutions

Notes on the Lebesgue Integral by Francis J. Narcowich November, 2013

Entrance Exam, Real Analysis September 1, 2017 Solve exactly 6 out of the 8 problems

ABSTRACT INTEGRATION CHAPTER ONE

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

MATH 5616H INTRODUCTION TO ANALYSIS II SAMPLE FINAL EXAM: SOLUTIONS

REAL ANALYSIS I Spring 2016 Product Measures

Math 172 Problem Set 5 Solutions

Partial Solutions to Folland s Real Analysis: Part I

Review of measure theory

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

6.2 Fubini s Theorem. (µ ν)(c) = f C (x) dµ(x). (6.2) Proof. Note that (X Y, A B, µ ν) must be σ-finite as well, so that.

MTH 404: Measure and Integration

consists of two disjoint copies of X n, each scaled down by 1,

ANALYSIS QUALIFYING EXAM FALL 2016: SOLUTIONS. = lim. F n

Probability and Measure

Mathematical Methods for Physics and Engineering

Real Analysis Chapter 3 Solutions Jonathan Conder. ν(f n ) = lim

Final Exam Practice Problems Math 428, Spring 2017

Math 4121 Spring 2012 Weaver. Measure Theory. 1. σ-algebras

Evaluation of integrals

Compendium and Solutions to exercises TMA4225 Foundation of analysis

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

REAL ANALYSIS I HOMEWORK 4

Notes on the Lebesgue Integral by Francis J. Narcowich Septemmber, 2014

Chapter 4. Measure Theory. 1. Measure Spaces

MATH MEASURE THEORY AND FOURIER ANALYSIS. Contents

REAL ANALYSIS ANALYSIS NOTES. 0: Some basics. Notes by Eamon Quinlan. Liminfs and Limsups

x 0 + f(x), exist as extended real numbers. Show that f is upper semicontinuous This shows ( ɛ, ɛ) B α. Thus

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

1 Review of di erential calculus

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

Real Analysis Problems

Defining the Integral

Math 5051 Measure Theory and Functional Analysis I Homework Assignment 2

+ 2x sin x. f(b i ) f(a i ) < ɛ. i=1. i=1

MATH 312 Section 2.4: Exact Differential Equations

Lebesgue s Differentiation Theorem via Maximal Functions

Course 212: Academic Year Section 1: Metric Spaces

Problem set 1, Real Analysis I, Spring, 2015.

Lebesgue measure and integration

MEASURE AND INTEGRATION: LECTURE 18

Chapter 6. Integration. 1. Integrals of Nonnegative Functions. a j µ(e j ) (ca j )µ(e j ) = c X. and ψ =

Solutions Final Exam May. 14, 2014

HOMEWORK 7 SOLUTIONS

HOMEWORK ASSIGNMENT 5

Measure and integration

Product measures, Tonelli s and Fubini s theorems For use in MAT4410, autumn 2017 Nadia S. Larsen. 17 November 2017.

4 Integration 4.1 Integration of non-negative simple functions

3 hours UNIVERSITY OF MANCHESTER. 22nd May and. Electronic calculators may be used, provided that they cannot store text.

Exercise 8.1 We have. the function is differentiable, with. f (x 0, y 0 )(u, v) = (2ax 0 + 2by 0 )u + (2bx 0 + 2cy 0 )v.

MATH 6337 Second Midterm April 1, 2014

THEOREMS, ETC., FOR MATH 516

Annalee Gomm Math 714: Assignment #2

A note on some approximation theorems in measure theory

Real Analysis Prelim Questions Day 1 August 27, 2013

Summer Jump-Start Program for Analysis, 2012 Song-Ying Li

Solutions: Problem Set 4 Math 201B, Winter 2007

LEBESGUE INTEGRATION. Introduction

First Order Differential Equations

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

MATH41011/MATH61011: FOURIER SERIES AND LEBESGUE INTEGRATION. Extra Reading Material for Level 4 and Level 6

Continuity. Chapter 4

MATH 31BH Homework 1 Solutions

Identity. "At least one dog has fleas" is translated by an existential quantifier"

Real Analysis: Homework # 12 Fall Professor: Sinan Gunturk Fall Term 2008

Math212a1413 The Lebesgue integral.

Methods of Integration

Problem Set 2: Solutions Math 201A: Fall 2016

Chapter 1.6, Page 37 Problem 2: (a) Prove that x is in the Cantor set iff x has a ternary expansion that uses only 0 s and 2 s.

Economics 204 Fall 2011 Problem Set 2 Suggested Solutions

II - REAL ANALYSIS. This property gives us a way to extend the notion of content to finite unions of rectangles: we define

Folland: Real Analysis, Chapter 2 Sébastien Picard

REAL AND COMPLEX ANALYSIS

1 Joint and marginal distributions

Lecture 4: Completion of a Metric Space

2.2 Separable Equations

Math 172 Problem Set 8 Solutions

SOME PROBLEMS IN REAL ANALYSIS.

MATS113 ADVANCED MEASURE THEORY SPRING 2016

Thus f is continuous at x 0. Matthew Straughn Math 402 Homework 6

Continuous Functions on Metric Spaces

Transcription:

SOLUTIONS OF SELECTED PROBLEMS Problem 36, p. 63 If µ(e n < and χ En f in L, then f is a.e. equal to a characteristic function of a measurable set. Solution: By Corollary.3, there esists a subsequence χ Enj f a.e.. Let A = {x : χ Enj (x f(x}. For x A, {χ Enj (x} is a sequence of s and s, so its limit can be either or. Let E = {x A : f(x = }. Then f = χ E a.e.. Problem 38-b, p. 63 Suppose that µ(x <, f n, g n are measurable functions, and f n f, g n g in measure. Prove that f n g n fg in measure. Show that it may not be the case if µ(x =. Solution:. Notice that n {x : f(x > n} =. lim n µ({x : f(x > n} = (here, we have used µ(x <. the same is true for the function g. we conclude that for every η > there exists a number M such that ( µ({x : f(x > M} < η and µ({x : g(x > M} < η.. For sufficiently large values of n, one has µ({x : f(x f n (x } < η. Notice that {x : f n (x > M + } {x : f(x > M} {x : f(x f n (x }. The same is true if one replaces f by g. there exists N such that ( µ({x : f n (x > M +} < η and µ({x : g n (x > M +} < η, n > N. 3. Fix a number ɛ >. Notice that Then, and {x : f n (xg n (x f(xg(x ɛ} {x : g n (x(f n (x f(x ɛ/} {x : f(x(g n (x g(x ɛ/}. {x : g n (x(f n (x f(x ɛ/} {x : g n (x > M + } {x : f n (x f(x ɛ/((m + } {x : f(x(g n (x g(x ɛ/} {x : f(x > M} {x : g n (x g(x ɛ/(m}. Typeset by AMS-TEX

SOLUTIONS OF SELECTED PROBLEMS µ({x : f n (xg n (x f(xg(x ɛ} 4η + δ n, n > N where δ n = µ({x : f n (x f(x ɛ/((m +}+µ({x : g n (x g(x ɛ/(m} when n. Hence, µ({x : f n (xg n (x f(xg(x ɛ} 5η when n is large enough. This completes the proof (η was an arbitrary number. 4. Let us drop the condition µ(x <. Take X = R, µ = m- the Lebesgue measure, f n (x = g n (x = x + /n, f(x = g(x = x. Clearly, f n f and g n g in measure. On the other hand, (x + /n x ɛ when x nɛ/. Therefore µ({x : f n (xg n (x f(xg(x ɛ} = for every ɛ > and for every n. Remark. The assumption µ(x < was used in the proof when we derived (. The same conclusion holds in the case when both functions f(x and g(x are integrable. Then ( is a consequence of the Chebyshev inequality µ({x : f(x > M} M X f(x dµ. If f(x and g(x are integrable functions then the assumption µ(x < can be dropped. Problem 43-b, p. 63 Suppose that µ(x < and f : X [, ] C is a function such that f(, y is measurable for each y and f(x, is continuous for each x. Then for every ɛ > there exists a measurable set E X with µ(e < ɛ and f(, y converges to f(, uniformly on E c as y. The solution is a modification of the proof of Egoroff s Theorem. Let E n (k = <r</n,r Q {x : f(x, r f(x, k. The sets E n (k are measurable: they are countable unions of measurable sets. For a fixed value of k, the sets E n (k are decreasing as n, and their intersection is empty. Therefore lim µ(e n(k =. n Choose n k in such a way that µ(e nk (k < k ɛ, and take E = k= E n k (k.

SOLUTIONS OF SELECTED PROBLEMS 3 Problem 44, p. 64 If f : [a, b] C is Lebesgue measurable and ɛ >, there is a compact set E [a, b] such that m(e c < ɛ and f E is continuous. (Lusin s Theorem Solution: As in the problem 38-b, there exists M such that m({x : f(x > M} < ɛ/3. We denote A = {x : f(x > M}. Let { f(x, when f(x M f M (x = M, when f(x > M. Then f M (x L, and, by Theorem.6, there exists a sequence of continuous functions that converges to f M in L. By Corollary.3, a subsequence of that sequence converges to f M a.e.. We conclude that there exists a sequence of continuous functions f n (x on [a, b] that converges to f M (x a.e.. By Egoroff s Theorem, there exists a set A such that m(a < ɛ/3 and f n converges to f uniformly on A c. Take a compact set E (A A c such that m((a A c \ E < ɛ/3. Notice that f(x = f M (x when x E. The restriction f E is the limit of uniformly convergent sequence of continuous functions f n E, so it is continuous. Clearly, m(e c < ɛ. Remark. Lusin s theorem does not say that f(x is continuous on E. It says that the restriction of f to E is continuous. For example, a function {, when x is rational D(x =, when x is irrational is not continuous at any point. However, the restriction of D(x to the set of irrational numbers (and to any subset of the set of irrational numbers is continuous. One can prove that a continuous function on a compact set can be extended to a continuous function on the whole interval (this is a special case of the Tietze extension theorem. Lusin s theorem implies the following fact: for every measurable function f(x on [a, b] and for every ɛ > there exists a continuous function f ɛ on [a, b] such that m({x : f(x f ɛ (x} < ɛ. In words, every measurable function can be changed on a set of arbitrarily small measure to get a continuous function. Problem 46, p. 68 (a part Let X = Y = [, ], M = N = B [,], µ Lebesgue measure, ν-counting measure. If D = {(x, x} [, ] is the diagonal, what is µ ν(d? Solution: Let D j=a j B j, A j, B j B [,] be a covering of D by rectangular sets. Let J = {j : ν(b j < }, let B = j J B j, and let E = [, ] \ B. Notice that the set B is either finite or countable; therefore µ(e =. On the other hand, E j J A j, so there exists j J such that µ(a j >. We conclude that µ(a j ν(b j =, and µ(a j ν(b j = j= for every covering of D by rectangles. µ ν(d =.

4 SOLUTIONS OF SELECTED PROBLEMS Problem 55, p. 77 Let E = [, ] [, ]. Investigate the existence and equality of fdm, f(x, ydxdy, and E f(x, ydydx for the following f. a f(x, y = (x y (x + y. Solution: Let us investigate substitution x = yz to get f(x, ydxdy first. For y >, one makes a f(x, ydx = y The substitution z = /w leads to /y z ( + z dz. so ( One notice that z ( + z dz = w ( + w dw, f(x, ydx = y /y /y z ( + z dz >, z ( + z dz. so The equality f(x, y = f(y, x implies f(x, ydydx = f(x, ydxdy f(x, ydxdy <. f(x, ydxdy >. f(x, ydydx, and the function f(x, y is not integrable on E. Remark. Though it not necessary to do for solving the problem, one can actually evaluate f(x, ydxdy. The function (z /[y(z + ] is positive in the domain {(z, y : < y <, z /y}, so one can use Tonelli s theorem to obtain (see ( f(x, ydxdy = z ( + z dz dy /z y = z ( + z ln zdz = I.

A substitution w = /z leads to SOLUTIONS OF SELECTED PROBLEMS 5 z w ( + z ln zdz = ( + w ln wdw. I = x ( + x ln xdx. Now, let us take the branch of the function ln z in the complex plane with negative imaginary half-axis removed that equals ln x on the positive real half-axis; ln z = ln z + i arg z, where π/ < arg z < 3π/. Then ln( x = ln x + iπ for x >, and In other words, z x ( + z ln zdz = 4I + iπ ( + x dx = 4I I = 4 z ( + z ln zdz. The last integral equals lim R,ɛ I(R, ɛ where z (3 I(R, ɛ = C R,ɛ ( + z ln zdz; here C R,ɛ = [ R, ɛ] {z : z = ɛ, Imz } [ɛ, R] {z : z = R, Imz }. When R > and ɛ <, the integrand in (3 has one pole, z = i, that lies inside the contour. The residue at this pole equals d dz ((z (z + i ln z z=i = i. 4I = (πi i = π. Finally, f(x, ydxdy = I = π 4. b f(x, y = ( xy a, a >. Solution: The function f(x, y is non-negative, so by Tonelli s theorem the integral over E and the iterated integrals are equal to each other. To find out, under what conditions they are finite, consider, say, f(x, ydxdy. The substitution z = xy leads to g(y = f(x, ydx = ( y a ( ay, when a ; ln( y y, when a =.

6 SOLUTIONS OF SELECTED PROBLEMS The function g(y is continuous on [,, and it goes to as y when a. The integral g(ydy is finite when a >, i.e. a <. We conclude that the function f(x, y is integrable when < a <. c ( 3 x f(x, y =, when < y < x (/ ;, otherwise. Solution: One has f((/ + z, y = f((/ z, y, therefore f(x, ydx =, and f(x, ydxdy =. On the other hand, h(x = and h(x dx = ( f(x, ydy = x 3 x, ( x dx =. the function h(x is not integrable, and the function f(x, y is not integrable as well. Answer: f(x, ydxdy =, integrals f(x, ydydx = and f(x, ydm E do not exist.