Harmonic functions, Poisson kernels

Similar documents
INTEGRATION WORKSHOP 2003 COMPLEX ANALYSIS EXERCISES

Weierstrass and Hadamard products

NATIONAL UNIVERSITY OF SINGAPORE Department of Mathematics MA4247 Complex Analysis II Lecture Notes Part I

HARMONIC FUNCTIONS. x 2 + 2

carries the circle w 1 onto the circle z R and sends w = 0 to z = a. The function u(s(w)) is harmonic in the unit circle w 1 and we obtain

One-Dimensional Dynamics

Appendix 2 Complex Analysis

SOME PROPERTIES OF CONJUGATE HARMONIC FUNCTIONS IN A HALF-SPACE

Math 115 ( ) Yum-Tong Siu 1. Derivation of the Poisson Kernel by Fourier Series and Convolution

30 Poisson Integral Formulas

Part D. Complex Analysis

= 2 x y 2. (1)

Key to Complex Analysis Homework 1 Spring 2012 (Thanks to Da Zheng for providing the tex-file)

u(e iθ )K(e iθ,z)dθ, z (1). (126) h(e iθ )K(e iθ,z)dθ, z (1),

Aero III/IV Conformal Mapping

MORE NOTES FOR MATH 823, FALL 2007

are harmonic functions so by superposition

Topic 3 Notes Jeremy Orloff

CONVERGENCE OF THE FOURIER SERIES

L-FUNCTIONS AND THE RIEMANN HYPOTHESIS. 1 n s

Fritz Hörmann MATH 316: Complex Analysis Fall 2010 Solutions to exercise sheet 2

Asymptotics at regular singular points

Hartogs Theorem: separate analyticity implies joint Paul Garrett garrett/

MATH 417 Homework 2 Instructor: D. Cabrera Due June 23. v = e x sin y

11 COMPLEX ANALYSIS IN C. 1.1 Holomorphic Functions

INTEGRATION WORKSHOP 2004 COMPLEX ANALYSIS EXERCISES

Math 185 Homework Exercises II

3.6 Applications of Poisson s equation

MAT389 Fall 2016, Problem Set 2

MATH COURSE NOTES - CLASS MEETING # Introduction to PDEs, Fall 2011 Professor: Jared Speck

MATH COURSE NOTES - CLASS MEETING # Introduction to PDEs, Spring 2018 Professor: Jared Speck

COMPLEX ANALYSIS AND RIEMANN SURFACES

+ b n sin 2πnt. dt. However, we have also seen that, using Euler s Formula, trigonometric functions can be written in a complex exponential form,

Primes in arithmetic progressions

III.2. Analytic Functions

Rudin Real and Complex Analysis - Harmonic Functions

Plurisubharmonic Functions and Pseudoconvex Domains

8 Complex Representations of Functions

Re(z) = a, For example, 3 + 2i = = 13. The complex conjugate (or simply conjugate") of z = a + bi is the number z defined by

Introduction to The Dirichlet Space

Complex Analysis, Stein and Shakarchi Meromorphic Functions and the Logarithm

Synopsis of Complex Analysis. Ryan D. Reece

Poisson Image Editing

Notes 7 Analytic Continuation

Leplace s Equations. Analyzing the Analyticity of Analytic Analysis DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING. Engineering Math EECE

COMPLEX DIFFERENTIAL FORMS. 1. Complex 1-forms, the -operator and the Winding Number

Assignment 2 - Complex Analysis

Department of Mathematics, University of California, Berkeley. GRADUATE PRELIMINARY EXAMINATION, Part A Fall Semester 2016

MAT389 Fall 2016, Problem Set 4

Poisson Image Editing

Conditioned Brownian Motion, Hardy spaces, Square Functions

8 8 THE RIEMANN MAPPING THEOREM. 8.1 Simply Connected Surfaces

The uniformization theorem

(z 0 ) = lim. = lim. = f. Similarly along a vertical line, we fix x = x 0 and vary y. Setting z = x 0 + iy, we get. = lim. = i f

Elliptic Equations. Chapter Definitions. Contents. 4.2 Properties of Laplace s and Poisson s Equations

Math Homework 2

ζ(u) z du du Since γ does not pass through z, f is defined and continuous on [a, b]. Furthermore, for all t such that dζ

Second Midterm Exam Name: Practice Problems March 10, 2015

The shortest route between two truths in the real domain passes through the complex domain. J. Hadamard

Lecture 4. Properties of Logarithmic Function (Contd ) y Log z tan constant x. It follows that

PROBLEM SET 6. E [w] = 1 2 D. is the smallest for w = u, where u is the solution of the Neumann problem. u = 0 in D u = h (x) on D,

Mathematics of Physics and Engineering II: Homework problems

The purpose of this lecture is to present a few applications of conformal mappings in problems which arise in physics and engineering.

Complex Variables & Integral Transforms

On Range and Reflecting Functions About the Line y = mx

Chapter 9. Analytic Continuation. 9.1 Analytic Continuation. For every complex problem, there is a solution that is simple, neat, and wrong.

Complex Analysis Prelim Written Exam Spring 2015

ELEMENTARY APPLICATIONS OF FOURIER ANALYSIS

Math 417 Midterm Exam Solutions Friday, July 9, 2010

LAPLACE EQUATION. = 2 is call the Laplacian or del-square operator. In two dimensions, the expression of in rectangular and polar coordinates are

Spotlight on Laplace s Equation

Poisson summation and convergence of Fourier series. 1. Poisson summation

Homework 27. Homework 28. Homework 29. Homework 30. Prof. Girardi, Math 703, Fall 2012 Homework: Define f : C C and u, v : R 2 R by

ZEROS OF MAASS FORMS

Department of Mathematics

Math 411, Complex Analysis Definitions, Formulas and Theorems Winter y = sinα

3.2 Differentiability; Tangent planes; differentials

MA30056: Complex Analysis. Revision: Checklist & Previous Exam Questions I

Solutions for Problem Set #5 due October 17, 2003 Dustin Cartwright and Dylan Thurston

Mid Term-1 : Solutions to practice problems

. Then Φ is a diffeomorphism from (0, ) onto (0, 1)

Complex Functions (1A) Young Won Lim 2/22/14

Complex Analysis. Travis Dirle. December 4, 2016

Title Project Summary

1. The COMPLEX PLANE AND ELEMENTARY FUNCTIONS: Complex numbers; stereographic projection; simple and multiple connectivity, elementary functions.

Module 7: The Laplace Equation

f (n) (z 0 ) Theorem [Morera s Theorem] Suppose f is continuous on a domain U, and satisfies that for any closed curve γ in U, γ

Leplace s Equations. Analyzing the Analyticity of Analytic Analysis DEPARTMENT OF ELECTRICAL AND COMPUTER ENGINEERING. Engineering Math 16.

Part IB. Complex Analysis. Year

Here are brief notes about topics covered in class on complex numbers, focusing on what is not covered in the textbook.

PICARD S THEOREM STEFAN FRIEDL

UNIFORM APPROXIMATION BY b HARMONIC FUNCTIONS

Math 120 A Midterm 2 Solutions

Complex Analysis Qual Sheet

UNIVERSIDAD CARLOS III DE MADRID MATHEMATICS II EXERCISES (SOLUTIONS )

A related space that will play a distinguished role in our space is the Hardy space H (D)

Harmonic Analysis Homework 5

A functional model for commuting pairs of contractions and the symmetrized bidisc

Appendix on complex numbers:

Theorem [Mean Value Theorem for Harmonic Functions] Let u be harmonic on D(z 0, R). Then for any r (0, R), u(z 0 ) = 1 z z 0 r

Transcription:

June 17, 16 Harmonic functions, Poisson kernels Paul Garrett garrett@math.umn.edu http://www.math.umn.edu/ garrett/ [This document is http://www.math.umn.edu/ garrett/m/complex/notes 14-15/8c harmonic.pdf] 1. Mean-value propert. Poisson kernel for disk 3. Dirichlet problem for disk 4. Poisson kernel for upper half-plane The two-dimensional Euclidean Laplacian is x + Man of these ideas are not specific to two dimensions: the Laplacian on R n is [1] x 1 +... + x n A twice-continuousl-differentiable complex-valued function u on a non-empt open set U C R satisfing Laplace s equation u is also called harmonic. In some contexts, a harmonic function is understood to be real-valued. With the notation we have z 1 x i z z z 1 x + i z z 1 4 A holomorphic function u satisfies the Cauch-Riemann equation u/z, so ever holomorphic function is harmonic. Similarl, ever ever conjugate-holomorphic function is harmonic. for holomorphic f, the real and imaginar parts Refz 1 are harmonic, and real-valued. fz + fz Imfz 1 i fz fz Harmonic functions have a mean-value propert similar to holomorphic functions. This ields the Poisson formula, recovering interior values from boundar values, much as Cauch s formula does for holomorphic functions. The solution of the Dirichlet problem is a converse: ever function on the boundar of a disk arises as the boundar values of a harmonic function on the disk. We prove this for relativel nice functions on the boundar, but with adequate set-up the same can be proven for an distribution generalized function on the boundar. [1] Up to scalar multiples, the Laplacian is the unique second-order differential operator on R n that is translationinvariant, rotation-invariant, and annihilates constants. In fact, ever translation-invariant and rotation-invariant differential operator on R n is a polnomial in the Laplacian. In some natural non-euclidean spaces there are motioninvariant differential operators not expressible in terms of the corresponding Laplacian. 1

Paul Garrett: Harmonic functions, Poisson kernels June 17, 16 1. Mean-value propert among other features, in two dimensions harmonic functions form a useful, strictl larger class of functions including holomorphic functions. For example, harmonic functions still enjo a mean-value propert, as holomorphic functions do: [1..1] Theorem: Mean-value propert For harmonic u on a neighborhood of the closed unit disk, u 1 ue iθ dθ Proof: Consider the rotation-averaged function vz 1 ue iθ z dθ for z 1 Since the Laplacian is rotation-invariant, v is a rotation-invariant harmonic function. In polar coordinates, for rotation-invariant functions vz f z, the Laplacian is v x + f x + x x z f z + z f z 1 z f x z 3 f + x z f + 1 z f z 3 f + z f f + 1 z f The ordinar differential equation f + f /r on an interval, R is an equation of Euler tpe, meaning expressible in the form r f + Brf + Cf with constants B, C. In general, such equations are solved b letting fr r λ, substituting, dividing through b r λ, and solving the resulting indicial equation for λ: λλ 1 + Aλ + B Distinct roots λ 1, λ of the indicial equation produce linearl independent solutions r λ1 and r λ. However, as in the case at hand, a repeated root λ produces a second solution r λ log r. Here, the indicial equation is λ, so the general solution is a + b log r. When b, the solution a + b log r blows up as r +. Since f v u is finite, it must be that b. a rotation-invariant harmonic function on the disk is constant. its average over a circle is its central value, proving the mean-value propert for harmonic functions. /// [1..] Remark: One might worr about commutation of the Laplacian with the integration above. In the first place, it is clear that we must have this commutativit. Second, the best and most final argument for such is in terms of Gelfand-Pettis also called weak integrals of function-valued functions, rather than temporar elementar arguments. [1..3] Remark: The solutions a + b log r do indeed exhaust the possible solutions: given f + f /r on, R, we see r f is constant because r f r f + f r f /r + f r The class of harmonic functions includes useful non-holomorphic real-valued functions. For example, realvalued logarithms of absolute values of non-vanishing holomorphic functions are harmonic: log fz 1 log f + log f 1 holomorphic + anti-holomorphic

so is annihilated b 4 z z. Paul Garrett: Harmonic functions, Poisson kernels June 17, 16. Poisson s formula and kernel for the disk The mean-value propert will ield [..1] Corollar: Poisson s formula For u harmonic on a neighborhood of the closed unit disk z 1, u is expressible in terms of its boundar values on z 1 b uz 1 Up to scalars, the Poisson kernel function is ue iθ 1 z z e iθ dθ for z < 1 P e iθ, z 1 z z e iθ Proof: Pre-composition h f of a harmonic function h with a holomorphic function f ields a harmonic 1 z function. With ϕ z the linear fractional transformation given b matrix ϕ z, the mean-value z 1 propert for u ϕ z gives uz u ϕ z 1 u ϕ z e iθ dθ Linear fractional transformations stabilizing the unit disk map the unit circle to itself. Replace e iθ b e iθ ϕ 1 z e iθ uz u ϕ z 1 ue iθ dθ Computing the change of measure will ield the Poisson formula. This is computed b ie iθ θ θ θ eiθ ie iθ 1 ze iθ + izeiθ e iθ z 1 ze iθ ieiθ ize iθ + ize iθ ie iθ z 1 ze iθ ieiθ 1 z 1 ze iθ giving the asserted integral. θ θ 1 e iθ ie iθ 1 z 1 ze iθ 1 zeiθ e iθ z eiθ 1 z 1 ze iθ 1 z z e iθ /// 3. Dirichlet problem on the disk As a sort of converse to the existence of the Poisson formula and Poisson kernel, the Dirichlet problem on the closed unit disk in C R is posed b specifing of a continuous function f on the circle S 1 {z : z 1}, and asking for harmonic u on z < 1, extending to a continuous function on z 1, such that u S 1 f. That is, the following asserts that the collection of harmonic functions is a correct collection of functions in which to pose the problem of reconstituting the interior values of a function from its boundar values, if we are to have existence and uniqueness. In contrast, the collection of holomorphic functions on a disk is too small, since the boundar values have Fourier series with no negative terms see below. On another hand, an class of functions strictl larger than harmonic functions will include non-zero functions with boundar values identicall. 3

Paul Garrett: Harmonic functions, Poisson kernels June 17, 16 [3..1] Corollar: Given a continuous function f on the circle S 1 {z : z 1}, there is a unique harmonic function u on the open unit disk extending to a continuous function on the closed unit disk and u S 1 f. In particular, uz 1 fe iθ For real-valued f, the function u is also real-valued. Proof: This proof re-discovers the Poisson kernel. 1 z z e iθ dθ for z < 1 To skirt some secondar analtical complications, suppose that f is at least C 1, so that the partial sums of the Fourier series of f converge uniforml and absolutel pointwise to f, so where fe it n Z fn 1 Note that on z 1 with z e it we can write In these terms, e int fn e int lim N e inθ fe iθ dθ z n for n and z e it fn e int z n for n < and z e it fn e int 1 e inθ fe iθ dθ e int 1 fe iθ e inθ e int dθ 1 1 ze fe iθ iθ N+1 1 ze iθ + zeiθ ze iθ N+1 1 ze iθ dθ After moving z to z < 1, we can take the limit N, and fn z n + n< n fn z n 1 fe iθ 1 zeiθ + 1 ze iθ 1 ze iθ dθ The left-hand side is the sum of an anti-holomorphic and a holomorphic function, so is harmonic. Simplifing slightl, 1 zeiθ + 1 ze iθ 1 ze iθ 1 zeiθ + ze iθ z 1 ze iθ fn z n + n< n fn z n 1 1 z 1 ze iθ 1 z e iθ z 1 z z e iθ fe iθ 1 z z e iθ dθ To see that we recover the original boundar values, let z re it with r >, so lim r 1 1 fe iθ 1 z z e iθ dθ lim fn r n e int + fn r n e int r 1 n n< n fne int fe it using the good convergence of the Fourier series for f on the circle. /// 4

Paul Garrett: Harmonic functions, Poisson kernels June 17, 16 4. Poisson kernel for upper half-plane Again using the fact that h f is harmonic for h harmonic and f holomorphic, we can transport the Poisson kernel P e iθ, z for the disk to a Poisson kernel for the upper half-plane H via the Cale map C : z z + i/iz + 1. The Cale map gives a holomorphic isomorphism of the disk to the upper half-plane, and of the circle with i removed to the real line. In the relation uz 1 fe iθ P e iθ, z dθ replace z b C 1 z with z H, and replace e iθ b C 1 z / with t R: using dθ de iθ /ie iθ, u C 1 z 1 d f C 1 t P C 1 t, C 1 z R The change of measure can be determined b the natural computation t + δ i it + δ + 1 t + δ i it + 1 iδ t i it + 1 + δ 1 + iδ1 it 1 + Oδ it + 1 d t + δ i it + 11 iδ1 it 1 t + δ i it + 1 1 iδ1 it 1 1 t i it + 1 + δ 1 it + 1 + t i i it + 1 it + 1 t i it + 1 + δ it + 1 + Oδ dt dt 1 + it1 it dt 1 + t With z x + i in H, the Poisson kernel itself becomes P H t, x + i P disk C 1 t, C 1 z 1 C 1 z C 1 z C 1 t 1 z i iz+1 z i iz+1 1 it 1 iz 1 + iz z i1 it 1 izt i 1 it x + + 1 x + 1 z i itz t t itz i z with all these coordinate changes, u C 1 z 1 f C 1 t R 1 it 4 4 z t 1 it x t + x t + dt 1 f C 1 t π R x t + dt + Oδ That is, replacing f C 1 b f and u C 1 b u, for suitable function f on the real line, a harmonic function u on the upper half-plane with boundar value f on R is given b ux + i 1 ft π R x t dt Poisson formula on H + 5