Problem Set. Problems on Unordered Summation. Math 5323, Fall Februray 15, 2001 ANSWERS

Similar documents
(C) The rationals and the reals as linearly ordered sets. Contents. 1 The characterizing results

Theorem Let J and f be as in the previous theorem. Then for any w 0 Int(J), f(z) (z w 0 ) n+1

SEPARATED AND PROPER MORPHISMS

Supplementary material for Continuous-action planning for discounted infinite-horizon nonlinear optimal control with Lipschitz values

VALUATIVE CRITERIA FOR SEPARATED AND PROPER MORPHISMS

Defining the Integral

Math 216A. A gluing construction of Proj(S)

1. Summation. Let X be a set Finite summation. Suppose Y is a set and + : Y Y Y. is such that

SEPARATED AND PROPER MORPHISMS

U e = E (U\E) e E e + U\E e. (1.6)

VALUATIVE CRITERIA BRIAN OSSERMAN

EXPANSIONS IN NON-INTEGER BASES: LOWER, MIDDLE AND TOP ORDERS

( x) f = where P and Q are polynomials.

Roberto s Notes on Differential Calculus Chapter 8: Graphical analysis Section 1. Extreme points

Maximum Flow. Reading: CLRS Chapter 26. CSE 6331 Algorithms Steve Lai

Math 421, Homework #7 Solutions. We can then us the triangle inequality to find for k N that (x k + y k ) (L + M) = (x k L) + (y k M) x k L + y k M

M17 MAT25-21 HOMEWORK 6

MATH 324 Summer 2011 Elementary Number Theory. Notes on Mathematical Induction. Recall the following axiom for the set of integers.

Discrete Mathematics. On the number of graphs with a given endomorphism monoid

HW 4 SOLUTIONS. , x + x x 1 ) 2

Climbing an Infinite Ladder

Finite Dimensional Hilbert Spaces are Complete for Dagger Compact Closed Categories (Extended Abstract)

On a Closed Formula for the Derivatives of e f(x) and Related Financial Applications

STAT 801: Mathematical Statistics. Hypothesis Testing

The 2nd Texas A&M at Galveston Mathematics Olympiad. September 24, Problems & Solutions

Math 324 Summer 2012 Elementary Number Theory Notes on Mathematical Induction

Maths 212: Homework Solutions

MATH 131A: REAL ANALYSIS (BIG IDEAS)

arxiv: v1 [math.oc] 1 Apr 2012

CHAPTER 8: EXPLORING R

Homework #2 Solutions Due: September 5, for all n N n 3 = n2 (n + 1) 2 4

Section 2.5 : The Completeness Axiom in R

Classification of effective GKM graphs with combinatorial type K 4

On High-Rate Cryptographic Compression Functions

Real Analysis Notes. Thomas Goller

CHOW S LEMMA. Matthew Emerton

Stat 451: Solutions to Assignment #1

9.3 Graphing Functions by Plotting Points, The Domain and Range of Functions

FIXED POINTS OF RENORMALIZATION.

EEO 401 Digital Signal Processing Prof. Mark Fowler

Homework 11. Solutions

Lebesgue-Radon-Nikodym Theorem

Published in the American Economic Review Volume 102, Issue 1, February 2012, pages doi: /aer

THEOREMS, ETC., FOR MATH 515

Midterm Exam Solutions February 27, 2009

MATH FINAL EXAM REVIEW HINTS

Indeed, if we want m to be compatible with taking limits, it should be countably additive, meaning that ( )

Homework 1 (revised) Solutions

THE CAUCHY PROBLEM VIA THE METHOD OF CHARACTERISTICS

In the index (pages ), reduce all page numbers by 2.

Math 754 Chapter III: Fiber bundles. Classifying spaces. Applications

6 CARDINALITY OF SETS

Math 2412 Activity 1(Due by EOC Sep. 17)

Journal of Mathematical Analysis and Applications

Introduction to Real Analysis Alternative Chapter 1

Special types of Riemann sums

GENERALIZED ABSTRACT NONSENSE: CATEGORY THEORY AND ADJUNCTIONS

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

Characterisation of Accumulation Points. Convergence in Metric Spaces. Characterisation of Closed Sets. Characterisation of Closed Sets

Chapter 4. Measure Theory. 1. Measure Spaces

Example: When describing where a function is increasing, decreasing or constant we use the x- axis values.

Descent on the étale site Wouter Zomervrucht, October 14, 2014

REAL ANALYSIS LECTURE NOTES: 1.4 OUTER MEASURE

Categories and Natural Transformations

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

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

EXISTENCE OF ISOPERIMETRIC SETS WITH DENSITIES CONVERGING FROM BELOW ON R N. 1. Introduction

2. The Concept of Convergence: Ultrafilters and Nets

Telescoping Decomposition Method for Solving First Order Nonlinear Differential Equations

h(x) lim H(x) = lim Since h is nondecreasing then h(x) 0 for all x, and if h is discontinuous at a point x then H(x) > 0. Denote

Mathematical Induction. Section 5.1

Ma 4121: Introduction to Lebesgue Integration Solutions to Homework Assignment 6

Thus, X is connected by Problem 4. Case 3: X = (a, b]. This case is analogous to Case 2. Case 4: X = (a, b). Choose ε < b a

Problem Set 2: Solutions Math 201A: Fall 2016

Supremum and Infimum

Fluctuationlessness Theorem and its Application to Boundary Value Problems of ODEs

CONSTRUCTION OF sequence of rational approximations to sets of rational approximating sequences, all with the same tail behaviour Definition 1.

CS 361 Meeting 28 11/14/18

Problem Set. Problem Set #1. Math 5322, Fall March 4, 2002 ANSWERS

ON MÜNTZ RATIONAL APPROXIMATION IN MULTIVARIABLES

Basic properties of limits

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

The basics of frame theory

Functions and cardinality (solutions) sections A and F TA: Clive Newstead 6 th May 2014

Sequences. Chapter 3. n + 1 3n + 2 sin n n. 3. lim (ln(n + 1) ln n) 1. lim. 2. lim. 4. lim (1 + n)1/n. Answers: 1. 1/3; 2. 0; 3. 0; 4. 1.

Outline. Approximate sampling theorem (AST) recall Lecture 1. P. L. Butzer, G. Schmeisser, R. L. Stens

The Clifford algebra and the Chevalley map - a computational approach (detailed version 1 ) Darij Grinberg Version 0.6 (3 June 2016). Not proofread!

means is a subset of. So we say A B for sets A and B if x A we have x B holds. BY CONTRAST, a S means that a is a member of S.

convergence theorem in abstract set up. Our proof produces a positive integrable function required unlike other known

F (x) = P [X x[. DF1 F is nondecreasing. DF2 F is right-continuous

Math 117: Topology of the Real Numbers

Extreme Values of Functions

CHAPTER I THE RIESZ REPRESENTATION THEOREM

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

Slowly Changing Function Oriented Growth Analysis of Differential Monomials and Differential Polynomials

Chapter 5. Measurable Functions

Finite-dimensional spaces. C n is the space of n-tuples x = (x 1,..., x n ) of complex numbers. It is a Hilbert space with the inner product

Lebesgue Integration: A non-rigorous introduction. What is wrong with Riemann integration?

On Picard value problem of some difference polynomials

Numerical Sequences and Series

Transcription:

Problem Set Problems on Unordered Summation Math 5323, Fall 2001 Februray 15, 2001 ANSWERS

i

1 Unordered Sums o Real Terms In calculus and real analysis, one deines the convergence o an ininite series (1.1) n=1 o reals numbers as ollows: For each postive integer n, deine n S n = a k, a n k=1 the n-th partial sum o the series. I the sequence { S n } converges to a real number S, the series converges and the sum is S. This deinition depends on the particular order o the terms a n i we reorder the terms, we get a dierent sequence o partial sums, and so possibly a dierent result. In this problem set, we want to consider how to add up ininitly many terms (including, perhaps, more than countably many terms) in a way that does not depend on the ordering. In our series (1.1) the sequence { a n } o terms amounts to a unction rom the index set N = { 1, 2, 3,... } to the real numbers R. In general we will consider an arbitary index set, and a unction : R. I F is a inite set, then the inite sum (x), x F makes sense, since the order o the terms does not matter in a inite sum. For brevity, we write = = (x). F F x F The emptyset is a inite set with = 0, or any unction. Note that i F 1, F 2 are inite and F 1 F 2 =, then = +, F 1 F 2 F F 2 since there are no duplicate terms in the sum on the let. Since we ll be using inite subsets o a lot, we introduce the notation Fin() = { F F is inite }. In the deinition o the sum o a series like (1.1), the idea is the we can add up initely many terms, and i we add in more and more terms we get closer and closer to the sum but we have a particular order in which to put in additional terms. In the ollowing deinition, we use the idea that adding in more terms should get closer to the sum, but we don t impose any order on which terms to take next. 1

Deinition 1.1. Let be a nonempty set and let : R. We say that is summable over i there is a real number S with the ollowing property: For every ε > 0, there is a inite set F 0 such that S < ε, whenever F F0 and F is inite. F Lemma 1.2. There is at most one number S that satisies the property in Deinition 1.1. Proo. Suppose that S 1 and S 2 have the property in the deintion. Let ε > 0 be arbitary. Then there is an F 1 Fin() such that F Fin(), F F 1 = S 1 < ε. F Similarily, there is an F 2 Fin() such that F Fin(), F F 2 = S 3 < ε. F I we set F = F 1 F 2 Fin(), then F F 1, F 2 and we have S 1 S 2 = S 1 + S 2 F F S 1 + F F S 2 < ε + ε = 2ε. Thus, we have S 1 S 2 < 2ε. Since ε was arbitary, we must have S 1 S 2 = 0, i.e., S 1 = S 2 I is summable, we will denote the unique number S in the deintion by or or (x), x and we introduce the notation S() = { : R is summable }. Problem 1. Suppose that : R. A. I itsel is inite, we ve given two deinitions o the symbol. Show these two deinitions are the same. B. I (x) = 0 or all x, then = 0. 2

Problem 2. The space S() o summable unctions is a vector space (under the pointwise operations) and the mapping S() R: is linear. In other words, A. I, g S(), then + g is summable and (1.2) ( + g) = + g. We want to show that the number + g satisies the condition in Deinition 1.1. Let ε > 0 be arbitrary. Since is summable, there is a inite set F 1 such that F Fin(), F F 1 = < ε. F Similarly, there is a inite set F 2 such that F Fin(), F F 2 = g g < ε. F Let F 3 = F 1 F 2, so F 3 is inite and F 3 F 1, F 2. Thus, i F F 3, we have [ ( + g) + ] g = + g g F F F F + g g F < ε + ε = 2ε. Since ε was arbitrary, this completes the proo. B. I S() and c R, c is summable and (1.3) c = c. We need to show the number on the right o (1.3) satisies the condition in Deinition 1.1. Let ε > 0 be arbitrary. Since is summable, there is a inite set F 0 such that F Fin(), F F 0 = < ε. F 3

Then, i F F 0, we have F c c = c F c ε. Since ε was arbitrary, c ε can be as small as we like, so this completes the proo. Problem 3. Let : R. A. The unction is summable over i and only i the ollowing Cauchy Condition is satisied: For every ε > 0, the is a inite set F 0 so that F, F Fin(), F, F F 0 = < ε. F F First suppose that is summable. We show that the Cauchy condition holds. Let ε > 0 be given. Then there is a inite set F 0 such that F Fin(), F F 0 = < ε. F Suppose that F, F are inite sets that contain F 0. Then, F = + F F F + F F < ε + ε = 2ε. Since ε was arbitrary, the Cauchy Condition holds. For the second part o the proo, suppose the Cauchy condition holds. We must show that is summable. Applying the Cauchy condition with ε = 1/n or n N gives us a sequence o inite sets F n so that (A) F, F Fin(), F, F F n = F 1 < F n. I we set E n = F 1 F 2 F n, then E 1 E 2 E 3 and E n F n. We claim that the sequence { E n } is a Cauchy sequence o real numbers. To see this, let ε > 0 be given. Choose N N so that 1/N < ε. I m, n N then E m, E n E N, so by (A), we have E n E m < 1/N < ε. Thus, { E n } is Cauchy. 4

Since a Cauchy sequence o real numbers converges, there is a number S so that E n S as n. We claim that S =. To see this, let ε > 0 be arbitrary, and choose N 1 N so that 1/N 1 < ε. Then, as above, (B) F, F Fin(), F, F E N1 = < ε. F F On the other hand, we can ind N 2 so that (C) n N 2 = S < ε. E n Let N = max(n 1, N 2 ). Suppose that F E N. Then we have F E N < ε, by (B) and we have E N S < ε, by (C). Thus, S = F F + S E N E N + S F E N E N < ε + ε = 2ε. Since ε was arbitrary, we conclude that is summable. B. The Cauchy Condition is equivalent to the ollowing Alternate Cauchy Condition: For all ε > 0 there is a inite set G such that F Fin(), F G = = < ε. F First suppose the Cauchy condition holds. Let ε > 0 be arbitrary. Then there is a inite set F 0 so that F, F Fin(), F, F F 0 = < ε. F F Suppose that H Fin() and that H F 0 =. Then both F 0 and H F 0 contain F 0, so we have = H H + F 0 F 0 = < ε. H F 0 F 0 Thus, the Alternate Cauchy Condition holds. For the second part o the proo, assume that Alternate Cauchy Condition holds, and we want to show the Cauchy Condition holds. To do this, let ε > 0 be given and let G be as in the Alternate Cauchy Condition. Suppose 5

that F is inite and F G. Then we can write F = G H, where H G =, so H < ε. Now suppose that F 1, F 2 G and write F i = G H i (i = 1, 2) as above. Then we have F 1 F 2 = G H 1 G H 2 = + G H 1 G H 2 = H 1 H 2 + H 1 H 2 < ε + ε = 2ε. Since ε was arbitrary, we conclude the Cauchy Condition holds. I A, we can consider the unction A : A R: a (a), the restriction o to A. I A is summable over A, we say is summable over A and write the sum as just A (instead o A A). Problem 4. Let : R and let A be a nonempty subset o. I is summable over, then is summable over A. We will use the Cauchy Condition to show that is summable over A. To show the Cauchy Condition holds on A, let ε > 0 be arbitrary. Let B = \ A, so is the disjoint union o A and B. Since is summable over, is satisies the Cauchy Condition on, so there is a set G Fin() so that (A) F 1, F 2 Fin(), F 1, F 2 G = < ε. F 1 F 2 The set G is the disjoint union o G A and G B. Suppose that H 1, H 2 Fin(A) and that H 1, H 2 G A. Then H 1 (G B) G and H 2 (G B) G. Thus, we have H 1 (G B) H 2 (G B) < ε, by (A). But H 1 (G B) H 2 (G B) = + H 1 G B H 2 G B = H 1 H 2 6

and so we have < ε. H 1 H 2 We conclude that satisies the Cauchy Condition on A, and so is summable on A. Problem 5. Let : R and suppose that is the disjoint union o two sets A and B. Then is summable over i and only i is summable over both A and B and, in this case, (1.4) = +. A B First suppose that is summable over both A and B. We want to show that is summable over and that (1.4) holds. Let ε > 0 be arbitrary. Since is summable over A, there is a set F A Fin(A) so that F Fin(A), F F A = < ε. F A Similarly, there is a set F B Fin(B) so that F Fin(B), F F B = < ε. F B Set F = F A F B. I F Fin() and F F, we can write F and the disjoint union o F A and F B, with F A F A and F B F B. Then we have [ + ] = F A B A B (F A) (F B) = + A B F A F B + A B F A < ε + ε = 2ε. F B We conclude that is summable over and (1.4) hold. For the second part o the proo, we assume that is summable over. But then, by the previous problem, is summable over A and B and hence (1.4) holds by the irst part o the proo. 7

The result o the last problem is easily extended inductively to the case where is partitioned into a inite number o subsets A 1, A 2..., A n. In this case is summable over i and only i it is summable over each o the A i and, i this is so, = A 1 + A 2 + + A n. A simple, but useul, corollary to the last problem is the ollowing. Corollary 1.3. Suppose that A and that : R is zero outside A. Then is summable over i and only i it is summable over A and, in this case, = A Problem 6. Let : R be summable. Then, or every ε > 0, there is a inite set F 0 such that A F 0 = < ε. A This diers rom the deinition o summablitity because A is allowed to be ininite. Let ε > 0 be given. Since is summable, there is a inite set F 0 such that (A) < ε F or all inite F F 0. Suppose that A is any subset o that contains F 0. Since is summable over, it is summable over A, so there is a inite set G 0 such that (B) < ε F A or all inite sets F with A F G 0. Let F = F 0 G 0. Then F A and F F 0, G 0. Thus, both (A) and (B) hold. Then we have A = + A F F A + F F This completes the proo. < ε + ε = 2ε. 8

Our next step is to extend one direction o Problem??? to arbitary partitions o. Recall that a partition P o is a collection o nonempty subsets o such that the elements o P are pairwise disjoint and P = P =. P P It ollows that every point o is in exactly one o the elements o P Problem 7. Let S() and let P be any partition o. Since is summable over every subset o, we can deine a unction g : P R: P g(p ) by g(p ) = P, Then, the unction g is summable over P and P g =. Let ε > 0 be given. By the last problem, there an F 0 Fin() so that (A) A F 0 = < ε. A Every point x is contained in exactly one element o P, call it P (x). Deine F 0 by F 0 = { P (x) x F 0 }. Then F 0 is a inite subcollection o P and certainly F 0 (the union o the elements o F) contains F 0. Suppose that F is a inite subcollection o P that contains F 0. Then F F 0 F 0. I we label the elements o F by F = { P 1, P 2,..., P k }, we have F g = P F g = g(p 1 ) + g(p 2 ) + + g(p k ) = P 1 + P 2 + + P k = P 1 P k = F. But, (A) implies that < ε, F 9

because F F 0. Thus, we have g < ε. F or all inite subcollections F that contain F 0. This completes the proo. Note that the converse o this theorem is alse. That is, it is possible that is summable over every element o P, so that g is deined, and that g is summable over P, but is not summable over. For an example like this, let = N and deine : N R by (n) = ( 1) n. Let P be the partition P = { { 1, 2 }, { 3, 4 }, { 5, 6 },... }. Then is summable over each element o P and g is summable over P, but is not summable over N. To see this, note that i was summable over N, it would be summable over every subset o N. In particular, it would be summable over { 2, 4, 6, 8,... }, which is not the case (why?). 2 Unordered Sums o Postive Terms Suppose that : [0, ]. I F Fin(), the inite sum F makes sense (it is, i takes the value at some point o F ). For unctions with non-negative values, the inite sums have two monotonicity properties. First, F 1, F 2 Fin(), F 1 F 2 = F 1 F 2. Secondly, i g : [0, ] is another unction, we have g, F Fin() = F F g. For any unction : [0, ], we deine = = (x) x by which may be. { = sup } F Fin, F Problem 8. Suppose that, g : R. 10

A. I A and B are subsets o, (2.1) A B = A B By deinition, (A) (B) { = sup } F A, F inite A F { = sup } F B, F inite. B F The set on the right-hand side o (B) is larger than the set in (A), and so has a larger sup. This establishes (2.1). B. For any subset A o, (2.2) g = g. A A Suppose g. For any inite set F A, we have F g g. F A Thus, g is an upper bound or the set A { F F A, F inite }, and hence is bigger than or equal to the sup o this set, which is. A Thus, (2.2) holds. Problem 9. In this problem, we establish that A. I, g : [0, ] then is a close to linear as it can be. ( + g) = + g. 11

I F is a inite set, then F ( + g) = + + g. F F Taking the sup over F, we conclude that ( + g) + g. To complete the proo, it will suice to show the opposite inequality (A) + g ( + g) I ( + g) =, this inequality is true, so we are reduced to considering the case (( + g)) <. Since, g + g we have, g ( + g) <. Let ε > 0 be arbitrary. Then there is a inite set F 1 so that ε < F 1. Similarly, there is a inite set F 2 so that Let F = F 1 F 2. Then we have g ε < F 2 g. (2.3) ε < F 1 F (2.4) g ε < g g F 2 F Adding these inequalities (ignoring the middle term) gives + g 2ε < + g = ( + g) ( + g). F F F Thus, we have Since ε was arbitrary, this implies (A) + g 2ε < ( + g). 12

B. I c [0, ) and : [0, ], then c = c where we re using the deinition 0 = 0. I c = 0, both sides o the equation in the problem are zero. So consider the case c > 0. For inite F, we have F c = c F (which is the distributive law). Thus, we have { } { F c F Fin = c } { } F Fin() = c F F F Fin(),, rom which the equation in the problem ollows by an elementary property o sup s. Problem 10. A. Suppose that A and that : [0, ] is zero outside o A. Then =. A Since A, we have A, so it will suice to establish the opposite inequality. Suppose that F is inite. Since = 0 outside A, we have F = F A. Thus, Sup ing over F yields and the proo is complete. F =. F A A B. Suppose that : [0, ]. Let A be a subset o and deine χ A, the characteristic unction o A, by { 1, x A χ A (x) = 0, x / A. A 13

Then, we have A = χ A. Since χ A is zero outside A, so A χ A = A. χ A = A χ A. But, = χ A on A, C. Let : [0, ] and suppose that is the disjoint union o A and B. Then = +. A B This result is easily extended to a partition o into initely many subsets. Since is the disjoint union o A and B, χ A +χ B = 1. Thus, = χ A +χ B and so = χ A + χ B. Now apply the previous part to the problem. Problem 11. Let : [0, ] and let P be a partition o. Deine g : P [0, ] by g(p ) = P. Then P g =. I F = { P 1, P 2,... P k } is a inite subcollection o P, then g = g(p 1) + g(p 2 ) + + g(p k ) F = + +... P 1 P 2 = P 1 P 2 P k = F. P k Thus, i F is a inite subcollection o P, g =. F F 14

Taking the sup over inite subcollection F P, we conclude that P g. To get the opposite inequality, let F be a inite subset o and set F = { P (x) x F } where P (x) is the unique element o P containing x. Then F is a inite subcollection o P and F F. Thus, F F = g g. F P Supping over F gives and the proo is complete. g, P Problem 12. Let : [0, ] and suppose that is countably ininite. Let { x n } n=1 be any enumeration o. Then (2.5) = (x n ), n=1 where the right-hand side is the limit o the partial sums, which may be ininity. Let F n = { x 1, x 2,..., x n }. Then, s n = n (x k ), k=1 the nth partial sum o the series on the right o (2.5) is the same as F n. Thus, s n = F n. Since the partial sums s n converge upward to the sum o the series s, we conclude that s. For the opposite inequality, suppose that F is inite. Find the maximum value o n such that x n F. Say this maximum value is N. Then F F N, so = s N s. F F N Supping over F gives s and the proo is complete. 15

I : [0, ], we deine the carrier o, denoted carr(), by From our previous results, we have carr() = { x (x) 0 }. = carr(). I carr() is inite, the right hand side o this equation would be a inite sum. I carr() is countably ininite, we can apply the last problem. Problem 13. Suppose that : [0, ] and that Then carr() is countable. <. Let E n = { x (x) 1/n } or n N. Then carr() = (To see this, note is (x) > 0 then (x) 1/n or some n.) We claim that each o the sets E n is inite. To prove this, suppose, or a contradiction, that E N is ininite. Then, or any k N, we can ind a subset F k o E N consisting o k elements. We then have n=1 E n k N F k Thus, k/n. Since k is arbitrary, this would imply =, contrary to our hypothesis. We conclude that each E n is inite. Then carr() is a countable union o countable sets, and hence countable. 16