Sequences and Series

Similar documents
The Limit of a Sequence (Brief Summary) 1

Math 312, Intro. to Real Analysis: Homework #4 Solutions

Lecture 9: The law of large numbers and central limit theorem

Solutions to Problem Sheet 1

11.7 (TAYLOR SERIES) NAME: SOLUTIONS 31 July 2018

1 Estimating sensitivities

5. Best Unbiased Estimators

NORMALIZATION OF BEURLING GENERALIZED PRIMES WITH RIEMANN HYPOTHESIS

Limits of sequences. Contents 1. Introduction 2 2. Some notation for sequences The behaviour of infinite sequences 3

SELECTING THE NUMBER OF CHANGE-POINTS IN SEGMENTED LINE REGRESSION

Introduction to Probability and Statistics Chapter 7

EXERCISE - BINOMIAL THEOREM

Fourier Transform in L p (R) Spaces, p 1

Chapter 8. Confidence Interval Estimation. Copyright 2015, 2012, 2009 Pearson Education, Inc. Chapter 8, Slide 1

Rafa l Kulik and Marc Raimondo. University of Ottawa and University of Sydney. Supplementary material

Section Mathematical Induction and Section Strong Induction and Well-Ordering

A New Constructive Proof of Graham's Theorem and More New Classes of Functionally Complete Functions

Statistics for Economics & Business

SUPPLEMENTAL MATERIAL

Average Distance and Vertex-Connectivity

A point estimate is the value of a statistic that estimates the value of a parameter.

Topic-7. Large Sample Estimation

Summary. Recap. Last Lecture. .1 If you know MLE of θ, can you also know MLE of τ(θ) for any function τ?

EVEN NUMBERED EXERCISES IN CHAPTER 4

r i = a i + b i f b i = Cov[r i, f] The only parameters to be estimated for this model are a i 's, b i 's, σe 2 i

Sampling Distributions and Estimation

Further Pure 1 Revision Topic 5: Sums of Series

x satisfying all regularity conditions. Then

Monopoly vs. Competition in Light of Extraction Norms. Abstract

Monetary Economics: Problem Set #5 Solutions

MATH 205 HOMEWORK #1 OFFICIAL SOLUTION

18.S096 Problem Set 5 Fall 2013 Volatility Modeling Due Date: 10/29/2013

4.5 Generalized likelihood ratio test

1 The Black-Scholes model

0.1 Valuation Formula:

Asymptotics: Consistency and Delta Method

FINM6900 Finance Theory How Is Asymmetric Information Reflected in Asset Prices?

. The firm makes different types of furniture. Let x ( x1,..., x n. If the firm produces nothing it rents out the entire space and so has a profit of

We consider the planning of production over the infinite horizon in a system with timevarying

1 Basic Growth Models

Lecture 5: Sampling Distribution

STAT 135 Solutions to Homework 3: 30 points

Sampling Distributions and Estimation

point estimator a random variable (like P or X) whose values are used to estimate a population parameter

A random variable is a variable whose value is a numerical outcome of a random phenomenon.

CHAPTER 8 Estimating with Confidence

Standard Deviations for Normal Sampling Distributions are: For proportions For means _

INTERVAL GAMES. and player 2 selects 1, then player 2 would give player 1 a payoff of, 1) = 0.

Economic Computation and Economic Cybernetics Studies and Research, Issue 2/2016, Vol. 50

Notes on Expected Revenue from Auctions

physicsandmathstutor.com

Outline. Plotting discrete-time signals. Sampling Process. Discrete-Time Signal Representations Important D-T Signals Digital Signals

Unbiased estimators Estimators

14.30 Introduction to Statistical Methods in Economics Spring 2009

AMS Portfolio Theory and Capital Markets

Overlapping Generations

Diener and Diener and Walsh follow as special cases. In addition, by making. smooth, as numerically observed by Tian. Moreover, we propose the center

On Regret and Options - A Game Theoretic Approach for Option Pricing

An Empirical Study of the Behaviour of the Sample Kurtosis in Samples from Symmetric Stable Distributions

A DOUBLE INCREMENTAL AGGREGATED GRADIENT METHOD WITH LINEAR CONVERGENCE RATE FOR LARGE-SCALE OPTIMIZATION

Parametric Density Estimation: Maximum Likelihood Estimation

between 1 and 100. The teacher expected this task to take Guass several minutes to an hour to keep him busy but

These characteristics are expressed in terms of statistical properties which are estimated from the sample data.

arxiv: v5 [cs.ce] 3 Dec 2008

Random Sequences Using the Divisor Pairs Function

1 over the interval [1,2]. Ans: 458,999 16,128. Find the arc length of the graph of the function

The material in this chapter is motivated by Experiment 9.

A New Second-Order Corrector Interior-Point Algorithm for P (κ)-lcp

Problem Set 1a - Oligopoly

Exam 1 Spring 2015 Statistics for Applications 3/5/2015

Simulation Efficiency and an Introduction to Variance Reduction Methods

Combining imperfect data, and an introduction to data assimilation Ross Bannister, NCEO, September 2010

Marking Estimation of Petri Nets based on Partial Observation

Maximum Empirical Likelihood Estimation (MELE)

SCHOOL OF ACCOUNTING AND BUSINESS BSc. (APPLIED ACCOUNTING) GENERAL / SPECIAL DEGREE PROGRAMME

Inferential Statistics and Probability a Holistic Approach. Inference Process. Inference Process. Chapter 8 Slides. Maurice Geraghty,


ST 305: Exam 2 Fall 2014

Confidence Intervals. CI for a population mean (σ is known and n > 30 or the variable is normally distributed in the.

Chapter 5: Sequences and Series

Online appendices from Counterparty Risk and Credit Value Adjustment a continuing challenge for global financial markets by Jon Gregory

Lecture 4: Parameter Estimation and Confidence Intervals. GENOME 560 Doug Fowler, GS

ii. Interval estimation:

ON DIFFERENTIATION AND HARMONIC NUMBERS

ECON 5350 Class Notes Maximum Likelihood Estimation

We learned: $100 cash today is preferred over $100 a year from now

Binomial Model. Stock Price Dynamics. The Key Idea Riskless Hedge

5 Statistical Inference

ASYMPTOTIC MEAN SQUARE ERRORS OF VARIANCE ESTIMATORS FOR U-STATISTICS AND THEIR EDGEWORTH EXPANSIONS

Research Paper Number From Discrete to Continuous Time Finance: Weak Convergence of the Financial Gain Process

MA Lesson 11 Section 1.3. Solving Applied Problems with Linear Equations of one Variable

ON THE RATE OF CONVERGENCE

BASIC STATISTICS ECOE 1323

This paper provides a new portfolio selection rule. The objective is to minimize the

Chapter 8: Estimation of Mean & Proportion. Introduction

Hopscotch and Explicit difference method for solving Black-Scholes PDE

Policy Improvement for Repeated Zero-Sum Games with Asymmetric Information

Stochastic Processes and their Applications in Financial Pricing

Annual compounding, revisited

B = A x z

Transcription:

Sequeces ad Series Matt Rosezweig Cotets Sequeces ad Series. Sequeces.................................................. Series....................................................3 Rudi Chapter 3 Exercises........................................ 3.3. Exercise............................................. 3.3. Exercise............................................. 3.3.3 Exercise 3............................................. 4.3.4 Exercise 4............................................. 4.3.5 Exercise 6............................................. 4.3.6 Exercise 7............................................. 5.3.7 Exercise 8............................................. 5.3.8 Exercise 9............................................. 5.3.9 Exercise 0............................................ 6.3.0 Exercise............................................ 6.3. Exercise............................................ 7.3. Exercise 3............................................ 8.3.3 Exercise 4............................................ 9.3.4 Exercise 5............................................ 0.3.5 Exercise 6.............................................3.6 Exercise 7.............................................3.7 Exercise 8.............................................3.8 Exercise 9.............................................3.9 Exercise 0.............................................3.0 Exercise............................................ Sequeces ad Series. Sequeces Propositio.. For p > 0, lim p 0.. For p > 0, lim p. 3. lim. 4. For p > 0 ad α R, lim α (+p) 0. 5. For x <, lim x 0. Proof. I all the followig proofs, we assume that ɛ > 0 is give. For (), by the Archimedea property, we ca take ɛ p so that p (ɛ p ) p ɛ For (), fix p > 0. If p >, set x : p ad observe that x > 0. By the biomial theorem, p Hece, lim sup x lim first case by cosiderig p. For (3), set x : + x ( + x ) p x p 0. If p, the the result is trivial. If 0 < p <, the we retur to the. Observe that x 0 ad by the biomial theorem, ( + x ) ( ) x ( ) x x

By (), lim 0. For (4), fix α R ad p > 0, ad choose a positive iteger k > α. For > k, we have by the biomial theorem that ( ) ( + p) > p k! ( ) ( k + ) ( ) k p k ( k)!k! pk p k k > k! k! ( + p) < k k! k p k Hece, 0 < Sice k α > 0, k α 0 by (). For (5), take α 0 ad p x > 0 i (4).. Series α ( + p) < k k! k α p k The followig propositio attributed to Cauchy shows that the covergece of a series with mootoically decreasig terms is determied by the growth of a small subset of its terms. Propositio. (Cauchy s Covergece Test) Let ( ) be a decreasig sequece of real umbers bouded from below by 0. The the series coverges if ad oly if the series coverges. Proof. Suppose coverges, the k k a k ( a + k k ) a k a + ( k )a k, which implies the covergece of k k a k ad therefore k a k by the compariso test. If coverges, the k a k a + k+ j k + which implies that coverges by the compariso test. a j k a j, j k a k Cauchy s covergece test allows us to give a short proof of the covergece (ad divergece) coditios for the harmoic p-series. p Propositio 3. (Root Test) Let ( ) be a sequece of complex umbers, ad set α : lim sup. The. if α <, the series coverges;. if α >, the series diverges; 3. if α, the root test is icoclusive. Proof. Suppose α <. Sice fiitely may terms do ot affect the covergece of, we may assume that sup <. We see that is absolutely coverget by the compariso test, sice ( ) < α α α Aalogously, if α >, the there exists a subsequece of idices ( k ) k such that k coditio lim does ot hold. >, so the ecessary

To see that the case α provides us isufficiet iformatio to determie covergece or divergece, cosider the followig examples. If for all Z, the clearly diverges to, but for all. If, the lim sup ( ) lim, where the peultimate iequality follows from cotiuity, ad π 6. Propositio 4. (Ratio Test) For a sequece ( ) of complex umbers, the series a. coverges, if lim sup + < ;. diverges if a+ for all but fiitely may..3 Rudi Chapter 3 Exercises.3. Exercise We first prove the useful reverse triagle iequality. Lemma 5. For a, b C, a b a + b. Proof. Observe that a + b (a + b)(a + b) a + (ab + ba) + b a + Re(ab) + b Takig the square root of both sides completes the proof. a a b + b a b Let (s ) be a coverget sequece of complex umbers. For ɛ > 0, there exists N Z such that N implies s s. By the reverse triagle iequality, N s s s s < ɛ The coverse is false. Defie s ad s +. The (s ) oscillates betwee ad, but s for all..3. Exercise lim +. Proof. We ca write + + x, for x 0, so that + ( + x) + x + x x + x 0 x For ay 0 < c <, I claim that x c for all but fiitely may. Ideed, otherwise there exists a subsequee k such that k x + x k (c ) + c c + c < k 4 k 4 k for all k sufficietly large, which is a cotradictio. We see that Lettig c completes the proof. ( + ) + ( + x) + x ( c) 3

.3.3 Exercise 3 Set s :, ad for Z, defie s + : + s The the sequece (s ) coverges ad moreover, s < for all. Proof. By iductio, we see that s > > for all. Hece, s + + s s + < s + < I claim that s < s +. The base case follows from the mootoicity of the square root fuctio. Suppose the assertio holds for all j. The s + + s > + s s.3.4 Exercise 4 Defie a real sequece (s ) by The lim if s 0 ad lim sup s. Proof..3.5 Exercise 6 If : +, the for ay N Z, 0 s s : m m, m Z + s m m +, m Z so diverges to. N +, If : +, the usig our above estimate for +, we see that + < Hece, coverges by the compariso test. 3 Suppose : +z. I claim that the series coverges for z > ad diverges for z <. First, suppose z >. For ay N Z, + z z ( z ) k z z k z ( z ) z z which coverges as N by compariso with the geometric series. Now suppose that z <. 4

.3.6 Exercise 7 Suppose ( ) is a sequece of oegative real umbers such that coverges. The the series a coverges. Proof..3.7 Exercise 8 If coverges ad (b ) is a bouded, mootoically icreasig sequece, the b also coverges. Proof. By cosiderig real ad imagiary parts separately, it suffices to cosider the case where the are real. Without loss of geerality, we may assume that b > 0 for all. Set b : sup b. Let ɛ > 0 be give, ad choose N 0 Z such that N, M N 0 implies that N < ɛ. M+ sg( ) b M+ M+ b b ( N M+ ) < bɛ If 0 the b b sg( ) < b, ad if < 0, the sice b b, b b b b. Hece, ( N ) bɛ < b b We coclude that.3.8 Exercise 9 M+ M+ N M+ b < ɛ, so the series b coverges by the Cauchy criterio. The radius of covergece of 0 3 z is. Proof. Sice the fiite limit of products of sequeces is the product of the limits, we have that ( ) 3 lim sup( 3 ) lim sup ( ) 3 lim The radius of covergece of 0! z is ifiite. Proof. Applyig the biomial formula to, for large, we obtai the upper boud!! m! k!( k)! k!( k)! + k!( k)! m m e m + km+ k! + km+ km+ Sice e k k!, we ca choose N Z such that m, N implies that km+ k! < ɛ. Choose N > N such that N e implies that N < ɛ. The N! e N + kn+ k! k! k! < ɛ + ɛ ɛ We coclude that lim sup! 0, ad therefore the radius of covergece is ifiite. 5

The radius of covergece of 0 z is. Proof. Sice lim a 0 implies that lim a lim sup ( a, we have that ) lim ( ) (lim ) The radius of covergece 0 3 3 z is 3. Proof. By the same argumets used above, we have that lim sup ( 3 3 ) ( ) 3 lim (lim ) 3 3 3 3.3.9 Exercise 0 Let ( ) be a sequece of itegers, ifiitely may of which are ozero. The the radius of covergece of z is at most. Proof. Deote the radius of covergece of the series z by R [0, ]. If z coverges for some z C, the lim z 0 If z >, the for all sufficietly large, <. Sice Z, 0, for all but fiitely may idices, which cotradicts our hypothesis..3.0 Exercise Let ( ) be a sequece of positive real umbers. Suppose diverges. The + diverges. Proof. If + coverges, the a + 0, so for ɛ > 0 small, + < ɛ < ɛ + ɛ < ɛ ɛ If ( ) is ubouded, the ( ) is bouded by some positive costat M >. The a + does ot ted to 0 as, hece Hece, + diverges by the compariso test. If s : a + +, the > a ( + a ) + + a N ( + a N ) > + M + M + + does ot coverge. Suppose a N+ + + a N+k s N, s N+ s N+k s N+k k Z ad therefore s diverges. Proof. For ay k Z, s N+ < < s N+k. Hece, a N+ s N+ + + a N+k s N+k a N+ + + a N+k s N+k This lower boud shows that the series s s N+k s N s N+k s N s N+k diverges by the Cauchy criterio. 6

The sequece ( a s ) satisfies ad therefore the series s Proof. Sice s < s, we have that coverges. s s s s < s s s s s s s s Sice s ad the RHS i the above iequality is telescopig, we see that as N. The series ( ). s a s + [ ] a ( s s s + ) a s s N s + s + coverges. The covergece or divergece of Proof. The secod claim follows from otig that coverges by the compariso test. For the first claim, first cosider :. The + depeds o the sequece + ad that <. Hece, +, which diverges sice the harmoic series diverges. Now defie by { : m, m Z otherwise + I claim that diverges. Let α > 0. Sice diverges, we ca choose a iteger N sufficietly large so that M N implies that M > α. The But + coverges sice > N m m > α + N m,m Z + N m,m Z + m m + < May thaks to the math.stackexchage commuity for helpig me out with this last part..3. Exercise Suppose ( ) is a sequece of positive reals such that the series coverges. Defie tails If m <, the r : a m, m Z ad therefore r coverges. a m r m + + r > r r m 7

Proof. First, ote that r + < r ad sice coverges, r 0. Hece, a m + + > a m + + r m r + r + > r r m r r m r m r m r m r m where m <. Let ɛ > 0 ad m Z be give. Sice r 0, we ca choose > m sufficietly large so that r r m > ɛ, which implies that the series r does ot satisfy the Cauchy criterio. For Z, so that the series r coverges. r < ( r r + ), Proof. We ca write r r +, so that r ( r + r + ) ( r + r ) r + r r r sice r+ r <. For ay N Z, as N. Hece,.3. Exercise 3 r r < coverges. ( ) r+ ( + r ) r + r < ( r r + ), ( r r + ) ( r r N+ ) 0 If a k ad b k coverge absolutely, the the Cauchy product c k coverges absolutely ad ( ) ( ) c k a k b k Proof. Let ɛ (0, ) be give. Set A a k ad B b k. Choose M > 0 such that a k, b k M, choose m > 0 such that max {sup k a k, sup k b k } m, ad choose N 0 N such that { } N N 0 max a k, b k < ɛ kn+ kn+ The for N N 0, we have N AB ( ) ) ( ) N ( ) N c k A a k (B b k + A a k b k + B b k a k + ( < Mɛ + Mɛ + a i ) b j < 4Mɛ in jn To see absolute covergece, observe that for N N, ( k N c k a i b k i a i ) ( b j a i ) b j < i0 i j i j N<i+j N Sice the partial sums N c k are mootoically odecreasig, they coverge by the mootoe covergece theorem. It follows by iductio that the Cauchy product of a fiitely may absolutely coverget series a k,j, j r, is absolutely coverget with limit a k,a k, a kr,r 0 k + +k r a i b j 8

.3.3 Exercise 4 Let (s ) 0 be a sequece of complex umbers. For Z 0, we defie the th arithmetric mea σ by If lim s s, the lim σ s. σ : s 0 + s + + s + Proof. Choose ɛ > 0, ad let N Z be sufficietly large such that N implies that s s < ɛ. The σ s s 0 + s + + s s + s 0 + s + + s s ( + ) + + s 0 s + + s s Choose N N such that N + N j0 s j s < ɛ. The N σ s j0 s j s + + jn+ s j s + N j0 s j s N + + ɛ < ɛ Note that a sequece (s ) eed ot coverge i order for its arithmetic meas (σ ) to coverge. Let s : ( ). The { 0 (mod ) σ + 0 (mod ) from which it is immediate that σ 0. We ca eve costruct a sequece (s ) satisfyig s > 0 for all, lim sup s, yet lim σ 0. Defie { k k, for some k Z 0 s : k ( k, k+ ) For a give Z 0, let k k() be the maximal iteger such that k, so that σ s k 0 + s + + s j0 j + j j k(k + ) + 4( (k+) ) + + ( + ) as. For, set : s s. The Proof. Observe that s σ s 0 + [s k s k ] + k s σ + s k s 0 + k k + k ka k k a k s 0 + + a k + a k + ka k k k k j a j ( k)a k k k(k + ) + 4 ( k + ) s 0 + k 0 k [s j s j ] j 9

Now assume that lim 0 ad lim σ σ exists. The lim s σ. Proof. Let ɛ > 0 be give. Choose N Z sufficietly large so that N implies that < ɛ ad σ σ < ɛ. Choose N N such that N implies that The + k a k < ɛ k N s σ s σ + σ σ < ɛ + + k ka k < ɛ + + < ɛ + + 3ɛ kn+ kn+ k a k ɛ It turs out that we ca relax our hypothesis above that lim 0 to just ( ) is a bouded sequece. Proof. Choose M > 0 such that < M for all. Let Z be large. For m <, write s σ m m s σ σ + m im+ ( m) i0 s i ( + )( m) s i + m im+ [s s i ] + ( + ) im+ s i + ( m)( + ) m im+ (m + )s i ( m) m i0 s i + ( m)( + ) m m + m σ m + m σ m + m im+ [s s i ] im+ im+ [s s i ] [s s i ] For i m +, s s i ki+ Choose ɛ > 0 small. The ɛ +ɛ a k ki+ a k ki+ M k ki+ M M( i) M( m ) i + i + m + lies i the iterval [m, m + ), for some oegative iteger m, so that m( + ɛ) ɛ < (m + )( + ɛ) (m + )ɛ m < + (m + )ɛ, which implies that m+ m ɛ ad therefore s s i < Mɛ, for m + i. Sice m as ad by our hypothesis that σ σ, we see that lim sup s σ lim sup Sice ɛ > 0 was arbitrary, we coclude that s σ..3.4 Exercise 5 [ σ σ m + m Mɛ + lim sup σ σ m Mɛ im+ For solutios to exercises 6, 7, ad 8, see my ote o square root algorithms. Mɛ ] 0

.3.5 Exercise 6.3.6 Exercise 7.3.7 Exercise 8.3.8 Exercise 9.3.9 Exercise 0 If (p ) is a Cauchy sequece i a metric space (X, d), ad some subsequece (p k ) k p X. The p p. coverges to a poit Proof. Choose N Z such that, m N implies that d(p, p m ) < ɛ. Choose k 0 Z such that k k 0 implies that k N ad d(p k, p) < ɛ. By the triagle iequality, d(p, p) d(p, p k ) + d(p k, p) ɛ + ɛ ɛ.3.0 Exercise For solutios to exercises 3, 4, ad 5, see my blog post etitled How to Complete Your Metric Space.