You are on page 1of 6

Module 9: The Method of Green’s Functions

The method of Green’s functions is an important technique for solving boundary value
and, initial and boundary value problems for partial differential equations.

In this module, we shall learn Green’s function method for finding the solutions of
partial differential equations. This is accomplished by constructing Green’s theorem that
are appropriate for the second order differential equations. These integral theorems will
then be used to show how BVP and IBVP can be solved in terms of appropriately defined
Green’s functions for these problems. More precisely, we shall study the construction and
use of Green’s functions for the Laplace, the Heat and the Wave equations.

1
MODULE 9: THE METHOD OF GREEN’S FUNCTIONS 2

Lecture 1 The Laplace Equation

Let Ω be a bounded domain in R2 . Consider the Laplace equation

∇2 u = 0 in Ω (1)

satisfying the BC
∂u
αu + β = B. (2)
∂n ∂Ω
Here, α(x), β(x), and B are given functions evaluated on the boundary region ∂Ω. The
∂u
term ∂n denotes the exterior normal derivative on ∂Ω. The boundary condition (2) relates
the values of u on ∂Ω and the flux of u through ∂Ω. We assume that α(x) > 0 and β(x) > 0
on ∂Ω.

If α ̸= 0, β = 0 then (2) is referred to as Dirichlet BC. If α = 0, β ̸= 0 then (2) is


referred to as Neumann BC. If α ̸= 0, β ̸= 0 then the condition (2) is known as Robin’s
type BC or mixed BC. We assume that ∂Ω is subdivided into three disjoint subsets ∂Ω1 ,
∂Ω2 and ∂Ω3 . On ∂Ω1 , ∂Ω2 and ∂Ω3 , u satisfies a boundary condition of the first kind
(Dirichlet type), second kind (Neumann type) and third kind (mixed type), respectively.

Introducing a function w(x) (whose property we shall specify later) and applying
Green’s theorem, we note that
∫ ∫ ( )
( ) ∂w ∂u
w∇ u − u∇ w dx =
2 2
u −w ds, (3)
Ω ∂Ω ∂n ∂n

where n is the exterior unit normal to ∂Ω. The equation (3) is the basic integral theorem
from which the Green’s function method proceeds in the elliptic case.

Now the function w(x) is to be determined so that (3) expresses u at an arbitrary


point ξ in the region Ω in terms of w and known functions in (1) and (2).

Let w(x) be a solution of


∇2 w = δ(x − ξ), (4)

where δ(x − ξ) is a two-dimensional Dirac delta function. Using the property of Dirac
delta function, we have
∫ ∫
2
u(∇ w)dx = u δ(x − ξ) dx = u(ξ). (5)
Ω Ω

In view of (1), we have ∫


w(∇2 u) dx = 0. (6)

MODULE 9: THE METHOD OF GREEN’S FUNCTIONS 3

It now remains to choose boundary conditions for w(x) on ∂Ω so that the boundary
integral in (3) involves only w(x) and known functions. This can be accomplished by
requiring w(x) to satisfy the homogeneous form of the given boundary condition (2), i.e.

∂w
αw + β = 0. (7)
∂n ∂Ω

If x ∈ ∂Ω1 on ∂Ω, we have


( )
∂w ∂u 1 ∂w ∂u
u −w = αu − αw
∂n ∂n α ∂n ∂n
( )
1 ∂w ∂u 1 ∂w
= αu + β = B , (8)
α ∂n ∂n α ∂n

where we have used (2). If x ∈ ∂Ω2 ∪ ∂Ω3 on ∂Ω, we have


( )
∂w ∂u 1 ∂u 1
u −w = − w αu + β = − Bw. (9)
∂n ∂n β ∂n β

The function w(x) is called the Green’s function for the boundary value problem (1)-(2).
To indicate its dependence on the point ξ, we denote the Green’s function by

w = G(x; ξ). (10)

In terms of G(x; ξ), (2) takes the form


∫ ∫
1 ∂G 1
u(ξ) = − B ds + BGds. (11)
∂Ω1 α ∂n ∂Ω2 ∪∂Ω3 β

The Green’s function G(x; ξ) thus satisfies the equation

∇2 G = δ(x − ξ), x, ξ ∈ Ω, (12)

and the BC
∂G
αG + β = 0. (13)
∂n ∂Ω

The Poisson equation in a rectangle: Let Ω : 0 < x < a, 0 < y < b be a rectangular
domain in R2 with boundary ∂Ω. Consider the following BVP

∇2 u = f (x, y) in Ω, (14)

with the BC

u(x, 0) = u(x, b) = 0, 0 < x < a, (15)


u(0, y) = u(a, y) = 0, 0 < y < b.
MODULE 9: THE METHOD OF GREEN’S FUNCTIONS 4

Let G(x, y; ξ, η) be the solution of the BVP

∇2 G = δ(x − ξ, y − δ) in Ω (16)

with the BC

G(x, 0; ξ, η) = G(x, b; ξ, η) = 0, 0 < x < a, (17)


G(0, y; ξ, η) = G(a, y; ξ, η) = 0, 0 < y < b,

where δ(x−ξ, y−η) = δ(x−ξ)δ(y−η). Since ∂Ω is piecewise smooth, we use the divergence
theorem to find that for a pair of smooth functions u and w.
∫ ∫
∂w ∂u
(w∇ u − u∇ w)dx =
2 2
(u − w )ds. (18)
Ω ∂Ω ∂n ∂n

Equation (18) is Green’s formula for functions of two space variables. If u is the solution
of the given BVP and w is replaced by G, then the homogeneous BC satisfied by both u
and G make the right-hand side in (18) vanish and the formula reduces to

[u(x, y)δ(x − ξ)δ(y − η) − f (x, y)G(x, y; ξ, η)]dxdy = 0. (19)

This yields ∫
u(ξ, η) = G(x, y; ξ, η)f (x, y)dxdy. (20)

Applying (18) with u(x, y) replaced by G(x, y; ξ, η) and w(x, y) replaced by G(x, y; ρ, σ),
we obtain
G(ξ, η; ρ, σ) = G(ρ, σ; ξ, η). (21)

Applying a simple interchange of variables, (20) becomes



u(x, y) = G(x, y; ξ, η)f (ξ, η)dξdη. (22)

G(x, y; ξ, η) is called the Green’s function of the given BVP. Formula (22) shows the effect
of all the sources in Ω on the temperature at the point (x, y).

To construct the Green’s function G, recall the two dimensional eigenvalue problem
associated with the BVP (14)-(15).

Uxx + Uyy + λU = 0, 0 < x < a, 0 < y < b,


U (0, y) = 0, U (a, y) = 0, 0 < y < b,
U (x, 0) = 0 U (x, b) = 0 0 < x < a.
MODULE 9: THE METHOD OF GREEN’S FUNCTIONS 5

The eigenvalues and the corresponding eigenfunctions are given by


( nπ )2 ( mπ )2 ( nπx ) ( mπy )
λnm = + , Unm = sin sin , n, m = 1, 2, . . .
a b a b

We now seek an expansion of the form


∞ ∑
∑ ∞
G(x, y; ξ, η) = cnm (ξ, η)Unm (x, y)
n=1 m=1
∑∞ ∑ ∞ ( nπx ) ( mπy )
= cnm (ξ, η) sin sin . (23)
a b
n=1 m=1

Putting (23) in (16), we obtain


∞ ∑
∑ ∞
∇2 G(x, y; ξ, η) = cnm (ξ, η)(∇2 Unm )(x, y)
n=1 m=1
∑∞ ∑∞
= − λnm cnm (ξ, η)Unm (x, y)
n=1 m=1
= δ(x − ξ)δ(y − η). (24)

Multiplying both sides of (24) by Upq . Then integrate over Ω and use the property of
Dirac delta function to obtain

4
cpq (ξ, η) = − Upq (ξ, η). (25)
abλpq

In view of (23), we now conclude


( ) ( )
4 ∑ ∞ sin nπξ sin mπη
∞ ∑
a b
( nπx ) ( mπy )
G(x, y; ξ, η) = − sin sin . (26)
ab ( nπ 2 mπ 2
a ) +( b ) a b
n=1 m=1

EXAMPLE 1. Using Green’s function method to solve

uxx + uyy = −π 2 sin(πx) sin(2πy), 0 < x < 1, 0 < y < 2,


u(x, 0) = 0, u(x, 2) = 0, 0 < x < 1,
u(0, y) = 0 u(1, y) = 0 0 < y < 2.

Here, a = 1, b = 2 and f (x, y) = −π 2 sin(πx) sin(2πy). By (23), we have


∞ ∑
∑ ∞
sin(nπξ) sin(mπη/2) ( mπy )
G(x, y; ξ, η) = −2 sin (nπx) sin .
(nπ)2 + (mπ)2 /4 2
n=1 m=1
MODULE 9: THE METHOD OF GREEN’S FUNCTIONS 6

It now follows from (22) that


∫ 2∫ 1 ∞ ∑
∑ ∞ ( mπy )
4 sin(nπξ) sin(mπη/2)
u(x, y) = (−2) sin (nπx) sin
0 0 π 2 (4n2 + m2 ) 2
n=1 m=1
×(−π ) sin(πξ) sin(2πη)dξdη
2

∑∞ ∑ ∞ (∫ 1 )
1
= 8 sin(πξ) sin(nπξ)dξ
4n2 + m2 0
n=1 m=1
(∫ 2 ( mπy ) ) ( mπy )
× sin(2πη) sin dη sin (nπx) sin
2 2
( 0 )
1 8 1
= sin(πx) sin(2πy) = sin(πx) sin(2πy).
2 4 · 12 + 42 5

REMARK 2. We know that separation of variables cannot be performed if the PDE and/or
BCs are not homogeneous. The eigenfunction expansion technique is used to deal with
IBVPs, where the PDE is nonhomogeneous and the BCs are zero.

Practice Problems

1. Use the Green’s function method to find the solution of the Dirichlet BVP:

uxx + uyy = x2 + y 2 , 0 < x < 1, 0 < y < 1,


u(x, 0) = u(x, 1) = 0, 0 < x < 1,
u(0, y) = u(1, y) = 0, 0 < y < 1.

2. Use the Green’s function method to find the solution of the Neumann BVP:

uxx + uyy = 0, 0 < x < 1, 0 < y < 1,


ux (x, 0) = ux (x, 1) = 0, 0 < x < 1,
u(0, y) = u(1, y) = 0, 0 < y < 1.

You might also like