Differentiating Series of Functions

Similar documents
Uniform Convergence and Series of Functions

General Power Series

Cable Convergence. James K. Peterson. May 7, Department of Biological Sciences and Department of Mathematical Sciences Clemson University

Uniform Convergence Examples

Uniform Convergence Examples

Geometric Series and the Ratio and Root Test

Integration and Differentiation Limit Interchange Theorems

Fourier Sin and Cos Series and Least Squares Convergence

Geometric Series and the Ratio and Root Test

Integration and Differentiation Limit Interchange Theorems

Notes on uniform convergence

Antiderivatives! Outline. James K. Peterson. January 28, Antiderivatives. Simple Fractional Power Antiderivatives

Dirchlet s Function and Limit and Continuity Arguments

Antiderivatives! James K. Peterson. January 28, Department of Biological Sciences and Department of Mathematical Sciences Clemson University

Linear Systems of ODE: Nullclines, Eigenvector lines and trajectories

Matrix Solutions to Linear Systems of ODEs

Hölder s and Minkowski s Inequality

Linear Systems of ODE: Nullclines, Eigenvector lines and trajectories

The Method of Laplace Transforms.

More Series Convergence

Fourier Sin and Cos Series and Least Squares Convergence

Mathematical Induction Again

Mathematical Induction Again

More Least Squares Convergence and ODEs

Lower semicontinuous and Convex Functions

Consequences of Continuity

Hölder s and Minkowski s Inequality

The Method of Undetermined Coefficients.

Derivatives and the Product Rule

Dirchlet s Function and Limit and Continuity Arguments

An idea how to solve some of the problems. diverges the same must hold for the original series. T 1 p T 1 p + 1 p 1 = 1. dt = lim

Convergence of Fourier Series

Riemann Sums. Outline. James K. Peterson. September 15, Riemann Sums. Riemann Sums In MatLab

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

Problem 3. Give an example of a sequence of continuous functions on a compact domain converging pointwise but not uniformly to a continuous function

8.8 Applications of Taylor Polynomials

Advanced Calculus Math 127B, Winter 2005 Solutions: Final. nx2 1 + n 2 x, g n(x) = n2 x

MATH 140B - HW 5 SOLUTIONS

Review of Power Series

Convergence of Sequences

Bolzano Weierstrass Theorems I

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

The First Derivative and Second Derivative Test

The First Derivative and Second Derivative Test

Mathematical Induction

Principles of Real Analysis I Fall VII. Sequences of Functions

Real Analysis Math 125A, Fall 2012 Sample Final Questions. x 1+x y. x y x. 2 (1+x 2 )(1+y 2 ) x y

Project One: C Bump functions

The Existence of the Riemann Integral

9. Series representation for analytic functions

Convergence of Sequences

We are going to discuss what it means for a sequence to converge in three stages: First, we define what it means for a sequence to converge to zero

Consequences of Continuity

Defining Exponential Functions and Exponential Derivatives and Integrals

2 2 + x =

Homework for MATH 4603 (Advanced Calculus I) Fall Homework 13: Due on Tuesday 15 December. Homework 12: Due on Tuesday 8 December

Math 180A. Lecture 16 Friday May 7 th. Expectation. Recall the three main probability density functions so far (1) Uniform (2) Exponential.

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

MT804 Analysis Homework II

First In-Class Exam Solutions Math 410, Professor David Levermore Monday, 1 October 2018

Solving systems of ODEs with Matlab

Taylor Polynomials. James K. Peterson. Department of Biological Sciences and Department of Mathematical Sciences Clemson University

Series. Definition. a 1 + a 2 + a 3 + is called an infinite series or just series. Denoted by. n=1

Math 104: Homework 7 solutions

Exercises from other sources REAL NUMBERS 2,...,

Measure and Integration: Solutions of CW2

Complex Analysis Slide 9: Power Series

Matrices and Vectors

R N Completeness and Compactness 1

Lecture Notes 3 Convergence (Chapter 5)

Solutions Final Exam May. 14, 2014

Derivatives in 2D. Outline. James K. Peterson. November 9, Derivatives in 2D! Chain Rule

MAT137 - Term 2, Week 2

The Limit Inferior and Limit Superior of a Sequence

Integration by Parts Logarithms and More Riemann Sums!

Math 328 Course Notes

Functions. Chapter Continuous Functions

APPROXIMATING CONTINUOUS FUNCTIONS: WEIERSTRASS, BERNSTEIN, AND RUNGE

Outline of Fourier Series: Math 201B

Analysis Comprehensive Exam, January 2011 Instructions: Do as many problems as you can. You should attempt to answer completely some questions in both

INTRODUCTION TO REAL ANALYSIS II MATH 4332 BLECHER NOTES

Homework 11. Solutions

Analysis Finite and Infinite Sets The Real Numbers The Cantor Set

The Comparison Test & Limit Comparison Test

Variation of Parameters

Some Background Material

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

Accumulation constants of iterated function systems with Bloch target domains

Functional Analysis HW #3

Predator - Prey Model Trajectories and the nonlinear conservation law

MATH 104 : Final Exam

Part IB Complex Analysis

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

Power Series. Part 1. J. Gonzalez-Zugasti, University of Massachusetts - Lowell

MA651 Topology. Lecture 9. Compactness 2.

Complex Analysis Qualifying Exam Solutions

Mathematics 805 Homework 9 Due Friday, April 3, 1 PM. j r. = (et 1)e tx t. e t 1. = t n! = xn 1. = x n 1

Lecture 4: Fourier Transforms.

Getting Started With The Predator - Prey Model: Nullclines

More Protein Synthesis and a Model for Protein Transcription Error Rates

Transcription:

Differentiating Series of Functions James K. Peterson Department of Biological Sciences and Department of Mathematical Sciences Clemson University October 30, 017

Outline 1 Differentiating Series

Differentiating Series Let s look at the problem of differentiating a series of functions. We start with an old friend, the series n=0 tn which we know converges pointwise to S(t) = 1/(1 t) on ( 1, 1). It also is known to converge uniformly on compact subsets of [ 1, 1]. We only showed the argument for sets like [ 1 + r, 1 r] but it is easy enough to go through the same steps for [c, d] in ( 1, 1). We used the uniform convergence to prove the pointwise limit had to continuous on the compact subsets of ( 1, 1) and easily extended the continuity to the whole interval ( 1, 1). This type of argument did not need to know the pointwise limit was 1/(1 t) and so we wokred through some other examples to illustrate how it works with series for which we cannot find the pointwise limit function. Our primary tools were the UCC which was a bit cumbersome to use, and the Second Weierstrass Uniform Convergence Theorem, which was much faster.

Differentiating Series Now let s look at what is called the derived series of n=0 tn. This is the series we get by differentiating the original series term by term. Let s call this series D(t) = n tn 1. Note the n = 0 term is gone as the derivative of a constant is zero. This is the series constructed from the partial sums S n. Using the ratio test, we find (n + 1) t n lim n n t n 1 (n + 1) = lim t = t n n Thus the series converges when t < 1 or on the interval ( 1, 1). At t = 1, the series is n which diverges and at t = 1, the series ( 1)n 1 n diverges by oscillation. To find out about uniform convergence, note on the interval [ 1 + r, 1 r] for sufficiently small r, we have t 1 r = ρ < 1. Define the sequence (K n ) by sup n t n 1 nρ n 1 = K n t [ 1+r,1 r]

Differentiating Series Then nρn 1 converges by the ratio test. We conclude by the Second Weierstrass Uniform Convergence Theorem that the series n tn 1 converges uniformly on [ 1 + r, 1 r] to a function V. Since each partial sum of n tn 1 is continuous, the uniform limit function V is continuous also and D = V on [ 1 + r, 1 r] since limits are unique. Finally, if t 0 in ( 1, 1) is arbitrary, it is inside some [ 1 + r, 1 r] and so V = D is continuous there also.

Differentiating Series Let s check the conditions of the derivative interchange theorem, applied to the sequence of partial sums (S n ). Fix a t in [0, 1). 1 S n is differentiable on [0, t]: True. S n is Riemann Integrable on [0, t]: True as each is a polynomial. 3 There is at least one point t 0 [0, t] such that the sequence (S n (t 0 )) converges. True as the series n=0 tn converges at t = 0. 4 S n unif y on [0, t] and the limit function y is continuous. True as we have just shown S n unif D on [0, t]. The conditions of derivative interchange theorem are satisfied and we can unif say there is a function W on [0, t] so that S n W on [0, t] and W = D. Since limits are unique, we then have W = S with S = D. We can do a similar argument for t ( 1.0]. Since we also know S(t) = 1/(1 t), this proves that on ( 1, 1) ( 1 1 t = ) ( ) 1 t n = nt n 1 1 = = 1 t (1 t) n=0

Example Examine convergence of the derived series for The derived series here is lim n 1 n+1 t n 1 n t n 1 = lim 5n ( n ) tn. 5 ( n ) tn 1. Using the ratio test, we find n 1 t = t /. This series thus converges when t / < 1 or t <. at t =, we have it sums to infintity. at t =, we have diverges by oscillation. 5 n 1 = n 10 which diverges since 5 ( ) n 1 = n ( 1)n 1 1 5 which

Hence, this series converges to a function D pointwise on (, ). Recall the original series also converged at t = but we do not have convergence of the derived series at this endpoint. To find out about uniform convergence, note on the interval [ + r, r] for sufficiently small r, we have t / 1 r = ρ < 1. Define the sequence (K n ) by sup t [ 1+r,1 r] 1 5 n t n 1 = sup t [ 1+r,1 r] 5 ( t /)n 1 ρ n 1 = K n Then ρn 1 converges by the ratio test. We conclude by the Second Weierstrass Uniform Convergence Theorem that the derived series 5 ( n ) tn 1 converges uniformly on [ + r, r] to a function V. Since each partial sum of 5 ( n ) tn 1 is continuous, the uniform limit function V is continuous also and D = V on [ + r, r] since limits are unique.

Finally, if t 0 in (, ) is arbitrary, it is inside some [ + r, r] and so V = D is continuous there also. Note we have shown S n unif D on [ + r, r] for all r with [ + r, r] (, ). Let s check the conditions of the derivative interchange theorem, applied to the sequence of partial sums (S n ). Fix a t in [0, ). 1 S n is differentiable on [0, t]: True. S n is Riemann Integrable on [0, t]: True as each is a polynomial. 3 There is at least one point t 0 [0, t] such that the sequence (S n (t 0 )) converges. True as the series 5n ( n ) tn converges at t = 0. 4 S n unif y on [0, t] and the limit function y is continuous. True as we have just shown S n unif D on [0, t].

The conditions of the derivative interchange theorem are satisfied and we unif can say there is a function W on [0, t] so that S n W on [0, t] and W = D. Since limits are unique, we then have W = S with S = D. We can do a similar argument for t (, 0]. Hence, on (.), ( S (t) = ) 5n ( n ) tn = 5 ( n ) tn 1 = D(t) This is why we say we can differentiate the series term by term!

Example Examine convergence of the derived series for 1 3 n t n. The derived series is n 3 n t n 1 Applying the ratio test, we find lim n n+1 3 n+1 t n n 3 n t n 1 = lim n n + 1 n t /3 = t /3. This series thus converges when t /3 < 1 or t < 3. at t = 3, we have n 3 3 n 1 = n comparison to 1 n. 1 3n which diverges by at t = 3, we have n 3 ( 3) n 1 = n ( 1)n 1 n 3 which diverges by oscillation. Hence, this series converges to a function D pointwise on ( 3, 3)

To find out about uniform convergence, note on the interval [ 3 + r, 3 r] for sufficiently small r, we have t /3 1 r = ρ < 1. Define the sequence (K n ) by sup n t/3 n 1 nρ n 1 = K n t [ 1+r,1 r] Then ρn 1 converges by the ratio test. We conclude by the Second Weierstrass Uniform Convergence Theorem that the derived series 5 t n 1 converges uniformly on n [ 3 + r, 3 r] to a function V. Since each partial sum of n(t/3)n 1 is continuous, the uniform limit function V is continuous also and D = V on [ 3 + r, 3 r] since limits are unique. Finally, if t 0 in ( 3, 3) is arbitrary, it is inside some [ 3 + r, 3 r] and so V = D is continuous there also. Note we have shown S n unif D on [ 3 + r, 3 r] for all r with [ 3 + r, 3 r] ( 3, 3).

Let s check the conditions of the derivative interchange theorem applied to the sequence of partial sums (S n ). Fix a t in [0, 3). 1 S n is differentiable on [0, t]: True. S n is Riemann Integrable on [0, t]: True as each is a polynomial. 3 There is at least one point t 0 [0, t] such that the sequence (S n (t 0 )) converges. True as the series 1 3 t n converges at n t = 0. 4 S n unif y on [0, t] and the limit function y is continuous. True as we have just shown S n unif D on [0, t]. The conditions of derivative interchange theorem are satisfied and we can unif say there is a function W on [0, t] so that S n W on [0, t] and W = D. Since limits are unique, we then have W = S with S = D. We can do a similar argument for t ( 3, 0].

Hence, on ( 3, 3) ( ) S (t) = (t/3) n = n=0 n(t/3) n 1 = D(t) The series n=0 (t/3)n is a geometric series and converges to 1/(1 (t/3)) on ( 3, 3), Hence, we can also say ( ) ( 1 ) = (t/3) n = 1 (t/3) n=0 n(t/3) n 1 1/3 = (1 (t/3))

We can take higher derivatives of series. The series obtained by differentiating the derived series term by term is called the second derived series. The third derived series is the series derived from the second derived series and so forth. It is easiest to explain how this works in an example. Example For the series 5 n=0 8(n+1) t n examine the first, second and third derived series for convergence. The first, second and third derived series are (First Derived Series:) This is (Second Derived Series:) This is n= (Third Derived Series:) This is n=3 5n 8(n+1) t n 1. 5(n 1)n 8(n+1) t n. 5(n )(n 1)n 8(n+1) t n 3.

We find all of these series converge on ( 1, 1) using the ratio test: we want all of these limits less than 1: (First Derived Series:) lim n 5(n+1) 8(n+) t n (Second Derived Series:) lim n 5n = lim 8(n+1) t n 1 n 5(n)(n+1) 8(n+) t n 1 (Third Derived Series:) lim n = lim 5(n 1)n 8(n+1) t n n 5(n 1)(n)(n+1) 8(n+) t n 5(n )(n 1)n 8(n+1) t n 3 = lim n (n + 1) 3 t = t n(n + ) n(n + 1) 3 t = t (n 1)n(n + ) (n 1)n(n + 1) 3 t = t (n )(n 1)(n + )

It is clear from these limits that all three series converge pointwise to a limit function on ( 1, 1). We could test if they converge at the endpoints but we will leave that to you. We check all three series for uniform convergence in the same way. On the interval [ 1 + r, 1 r] for small enough r, we have t 1 r = ρ < 1. We have these estimates: (First Derived Series:) 5n 8(n + 1) t n 1 5(n + 1) 8(n + 1) ρn 1 1 (n + 1) ρn 1 So K n = 1 (n+1) ρn 1. Since K n converges by the ratio test since 0 ρ < 1, we see the first derived series converges uniformly on [ 1 + r, 1 r].

(Second Derived Series:) n= 5(n 1)n 8(n + 1) t n n= 5(n + 1) 8(n + 1) ρn n= ρ n So K n = ρ n. Since K n converges by the ratio test since 0 ρ < 1, we see the second derived series converges uniformly on [ 1 + r, 1 r]. (Third Derived Series:) n=3 5(n )(n 1)n 8(n + 1) t n 3 n=3 5(n + 1) 3 8(n + 1) ρn 3 (n + 1)ρ n 3 n=3 So K n = (n + 1)ρ n 3. Since K n converges by the ratio test since 0 ρ < 1, we see the third derived series converges uniformly on [ 1 + r, 1 r].

By the Second Weierstrass Uniform Convergence Theorem, we see all these derived series converge uniformly. We have (First Derived Series:) S n (Second Derived Series:) S n (Third Derived Series:) S n unif D on [ 1 + r, 1 r]. unif E on [ 1 + r, 1 r]. unif F on [ 1 + r, 1 r]. Since the partial sums of these derived series are continuous and the convergence is uniform D, E and F are continuous on [ 1 + r, 1 r]. Further given any t ( 1, 1), t is in some [ 1 + r, 1 r] and so we know D, E and F are continuous on ( 1, 1). Now we check the derivative interchange for these derived series

(First Derived Series:) 1 Fix a t in [0, 1). S n is differentiable on [0, t]: True. S n is Riemann Integrable on [0, t]: True as each is a polynomial. 3 There is at least one point t 0 [0, t] such that the sequence (S n (t 0 )) converges. True as the original series converges at t = 0. 4 S n unif y on [0, t] and the limit function y is continuous. True as we have just shown S n unif D on [0, t]. The derivative interchange theorem conditions are satisfied and so unif there is a function W on [0, t] with S n W on [0, t] and W = D. Since limits are unique, we then have W = S with S = D. Do a similar argument for t ( 1.0]. So on ( 1, 1) ( S (t) = ) 5 8(n + 1) tn 1 = 5n 8(n + 1) tn 1 = D(t)

(Second Derived Series:) 1 Fix a t in [0, 1). S n is differentiable on [0, t]: True. S n is Riemann Integrable on [0, t]: True as each is a polynomial. 3 There is at least one point t 0 [0, t] such that the sequence (S n(t 0 )) converges. True as the derived series converges at t = 0. 4 S n unif y on [0, t] and the limit function y is continuous. True as we have just shown S n unif E on [0, t]. The derivative interchange theorem is satisfied and so there is X on [0, t] so that S n unif X on [0, t] and X = E. Limits are unique, so X = D with D = E. From the first derived series, we have S = D = E. We can do a similar argument for t ( 1.0].

Hence, on ( 1, 1) ( D (t) = ( S (t) = n=0 ) 5n 8(n + 1) tn 1 = 5 8(n + 1) tn 1 ) = = n= ( n= 5(n 1)n 8(n + 1) tn = E(t) 5n 8(n + 1) tn 1 ) 5(n 1)n 8(n + 1) tn = E(t)

(Third Derived Series:) 1 Fix a t in [0, 1). S n is differentiable on [0, t]: True. S n is Riemann Integrable on [0, t]: True as each is a polynomial. 3 There is at least one point t 0 [0, t] such that the sequence (S n (t 0 )) converges. True as the second derived series converges at t = 0. 4 S n unif y on [0, t] and the limit function y is continuous. True as we have just shown S n unif F on [0, t]. The conditions of derivative interchange theorem are satisfied and we can say there is a function Y on [0, t] so that S n unif Y on [0, t] and Y = F. Since limits are unique, we then have Y = E with E = F. Using the results from the first and second derived series, we have S = E = F. We can do a similar argument for t ( 1, 0].

Hence, on ( 1, 1) ( E (t) = n= ( S (t) = ( = n= n=0 ) 5(n 1)n 8(n + 1) tn = 5 8(n + 1) tn 1 5(n 1)n 8(n + 1) tn ) = n=3 ) ( = n=3 5(n )(n 1)n 8(n + 1) t n 3 = F (t) 5n 8(n + 1) tn 1 ) 5(n )(n 1)n 8(n + 1) t n 3 = F (t)

Homework 6 You need to follow the full arguments we have done for the examples in this Lecture for these problems. So lots of explaining! 6.1 Examine the convergence of the derived series of n=0 5 n4 n t n. 6. Examine the convergence of the derived series and the second derived series of n=0 6 n t n.