SlideShare a Scribd company logo
1 of 73
Download to read offline
Subharmonic Functions and the Solution to the Dirichlet
Problem
Author: Bryan R. Clark
Advisor: Dr. Edoh Amiran
Western Washington University
brclark9@gmail.com
Thursday 26th September, 2013
1 / 73
Overview
1 Harmonic Functions
Properties of Harmonic Functions
Dirichlet Problem for the Disk
Harnack’s Principle
2 Subharmonic Functions
Properties
Barrier Function
Perr´on Method
2 / 73
Harmonic Functions
Definition
A real-valued function u : U → R on an open set U ⊆ C is harmonic if it
is C2 on U and
∆u ≡ 0
there where the Laplacian ∆u is defined by
∆u =
∂2
∂x2
+
∂2
∂y2
u.
3 / 73
Harmonic Functions
Example
Is u(x, y) = e3x cos(3y) harmonic?
4 / 73
Harmonic Functions
Example
Is u(x, y) = e3x cos(3y) harmonic?
∂u
∂x
= 3e3x cos(3y)
∂2u
∂x2
= 9e3x cos(3y)
∂u
∂y
= −3e3x sin(3y)
∂2u
∂y2
= −9e3x cos(3y)
Therefore ∆u = ∂2
∂x2 + ∂2
∂y2 u = 9e3x cos(3y) − 9e3x cos(3y) = 0 thus u
is harmonic.
5 / 73
Maximum/Minimum Principle
Theorem
If u : U → R is harmonic on a connected open set U and if there is a point
P0 ∈ U with the property that u(P0) = supQ∈U u(Q) or
u(P0) = infQ∈U u(Q), then u is constant on U.
6 / 73
Maximum/Minimum Principle
Theorem
If u : U → R is harmonic on a connected open set U and if there is a point
P0 ∈ U with the property that u(P0) = supQ∈U u(Q) or
u(P0) = infQ∈U u(Q), then u is constant on U.
An immediate consequence of this theorem is the following corollary.
Corollary
Let U ⊆ C be a bounded, connected open set and let u be a continuous,
real valued function on the closure U of U that is harmonic on U. Then
max u
U
= max u
∂U
min u
U
= min u
∂U
.
7 / 73
Maximum/Minimum Principle
Lets prove this corollary.
Proof.
A continuous function on a closed bounded set attains its maximum. Thus
there exists a P ∈ U such that u(P) is maximal. If P ∈ U then by the
maximum principle u is constant and thus attains its maximum. If P ∈ ∂U
then the result is certainly true.
8 / 73
Mean-value Property
Theorem
Suppose that u : U → R is a harmonic function on an open set U ⊆ C and
that D(P, r) ⊆ U for some r > 0 then
u(P) =
1
2π
2π
0
u(P + reiθ
)dθ.
9 / 73
Mean-value Property
Theorem
Suppose that u : U → R is a harmonic function on an open set U ⊆ C and
that D(P, r) ⊆ U for some r > 0 then
u(P) =
1
2π
2π
0
u(P + reiθ
)dθ.
Q: What does this mean?
10 / 73
Mean-value Property
A:
11 / 73
Mean-value Property
A:
Q: What does this remind you of?
12 / 73
Mean-value Property
A:
Q: What does this remind you of?
A: The Cauchy integral formula.
13 / 73
Mean-value Property
Example
Let u(x, y) = x2 − y2, harmonic, and the disk D(P, 1) where P = 1 + i.
Then
u(1, 1) =
1
2π
2π
0
u(1 + cos(θ), 1 + sin(θ)) dθ
=
1
2π
2π
0
[1 + cos(θ)]2
− [1 + sin(θ)]2
dθ
=
1
2π
2π
0
2 cos(θ) − 2 sin(θ) + cos(2θ) dθ
=
1
2π
2π
0
2 cos(θ) dθ −
1
2π
2π
0
2 sin(θ) dθ +
1
2π
2π
0
cos(2θ) dθ
= 0
14 / 73
Poisson Integral Formula
Theorem
Let u : U → R be a harmonic function on a neighborhood of D(0, 1).
Then, for any point a ∈ D(0, 1),
u(a) =
1
2π
2π
0
u(eiψ
)
1 − |a|2
|a − eiψ|2
dψ.
15 / 73
Poisson Integral Formula
Theorem
Let u : U → R be a harmonic function on a neighborhood of D(0, 1).
Then, for any point a ∈ D(0, 1),
u(a) =
1
2π
2π
0
u(eiψ
)
1 − |a|2
|a − eiψ|2
dψ.
The Poisson integral formula says that the value u(reiψ) for some interior
point reiψ is a weighted average of the values of u on the boundary of the
circle where the weight
P(θ − ψ) =
1
2π
·
1 − r2
1 + r2 − 2r cos(θ − ψ)
> 0
and has total weight 1.
16 / 73
Dirichlet Problem for the Disk
Definition
Let D denote the unit disk. Given a bounded, piecewise continuous
function on the boundary of the disk,
u : ∂D → R
is there a harmonic function on the disk,
u : D → R
that extends to match the original u on the boundary of the disk, the
extension begin continuous other than the original discontinuities?
17 / 73
Dirichlet problem for the Disk
Does there always exist a solution to the Dirichlet problem on the disk?
Yes or No
18 / 73
Solution to the Dirichlet Problem for the Disk
Theorem
Let f be a continuous function on ∂D(0, 1). Define
u(z) =



1
2π
2π
0 f (eiψ) · 1−|z|2
|z−eiψ|2 dψ if z ∈ D(0, 1)
f (z) if z ∈ ∂D(0, 1)
Then u is continuous on D(0, 1) and harmonic on D(0, 1).
19 / 73
Solution to the Dirichlet Problem for the Disk
Proof.
It can be shown that
1 − |z|2
|z − eiθ|2
=
eiθ
eiθ − z
+
e−iθ
e−iθ − ¯z
− 1.
When z ∈ D(0, 1),
u(z) =
1
2π
2π
0
f (eiθ
)
eiθ
eiθ − z
dθ +
1
2π
2π
0
f (eiθ
)
e−iθ
e−iθ − ¯z
dθ
−
1
2π
2π
0
f (eiθ
)dθ
20 / 73
Solution to the Dirichlet Problem for the Disk
Continued.
Using the fact that ∆ = 4 ∂2
∂z∂¯z we conclude that the first integral is
harmonic.
∂
∂¯z
1
2π
2π
0
f (eiθ
)
eiθ
eiθ − z
dθ =
1
2π
2π
0
f (eiθ
)
∂
∂¯z
eiθ
eiθ − z
dθ
= 0
21 / 73
Solution to the Dirichlet Problem for the Disk
Continued.
Using the fact that ∆ = 4 ∂2
∂z∂¯z we conclude that the first integral is
harmonic.
∂
∂¯z
1
2π
2π
0
f (eiθ
)
eiθ
eiθ − z
dθ =
1
2π
2π
0
f (eiθ
)
∂
∂¯z
eiθ
eiθ − z
dθ
= 0
Since eiθ
eiθ−z
is holomorphic. Thus it satisfies the criteria for being
harmonic. The second and third integrals are harmonic by the same
argument. So we have a sum of harmonic functions which is harmonic.
Therefore if z ∈ D(0, 1) then u is harmonic.
22 / 73
Solution to the Dirichlet Problem for the Disk
Continued.
Now I will convince you that the function u(z) is continuous.
When z ∈ D(0, 1) we want to show that u is continuous. The Poisson
integral formula says that the value u(reiθ) for some interior point reiθ is a
weighted average of the values of f on the boundary of the circle where
the weight P(θ − ψ) > 0 and has total weight 1. We see that if we fix θ
and take the limit limr→1− Pr (θ − ψ) we get 0 given θ = ψ. If r is very
close to 1, i.e. eiθ very close to the boundary, then Pr (θ − ψ) is close to 0
except when ψ is close to θ. We are taking a weighted average with most
of the weight concentrated near the point eiθ. Thus it makes sense that
u(reiθ) converges to f (eiθ).
23 / 73
Solution to the Dirichlet Problem for the Disk
Example
Find the function u harmonic on the unit disk, D(0, 1) = {z : |z| < 1} and
takes on the boundary vales f (t) =



π for π
2 < t < π
0 for − π
2 < t < π
2
−π for − π < t < −π
2
24 / 73
Solution to the Dirichlet Problem for the Disk
Example (Continued)
Solution:
u(z) =
− tan−1 1+r
1−r tan θ+π
2 + tan−1 1+r
1−r tan 2θ+π
4
π
+
tan−1 1+r
1−r tan 2θ−π
4 − tan−1 1+r
1−r tan θ−π
2
π
if z ∈ D(0, 1), and
u(z) = f (z)
if z ∈ ∂D(0, 1)
25 / 73
The Harnack Inequality
Proposition
Let u be a nonnegative, harmonic function on a neighborhood of D(P, R).
Then for any z ∈ D(P, R),
R − |z − P|
R + |z − p|
· u(P) ≤ u(z) ≤
R + |z − P|
R − |z − P|
· u(P).
26 / 73
The Harnack Inequality
Proposition
Let u be a nonnegative, harmonic function on a neighborhood of D(P, R).
Then for any z ∈ D(P, R),
R − |z − P|
R + |z − p|
· u(P) ≤ u(z) ≤
R + |z − P|
R − |z − P|
· u(P).
Harnack’s inequality is an inequality relating the values of a positive
harmonic function at two points.
27 / 73
The Harnack Inequality
Example
Let u(x, y) = 1
2(x + 1) and which is nonnegative on D(0, 1). Our upper
and lower bounds for a point z = 1
2 +
√
3
4 i are,
R − |z|
R + |z|
· u(0) ≤ u(z) ≤
R + |z|
R − |z|
· u(0).
1 − |1
2 +
√
3
4 |
1 + |1
2 +
√
3
4 |
· 0.5 u
1
2
+
√
3
4
i
1 + |1
2 +
√
3
4 |
1 − |1
2 +
√
3
4 |
· 0.5.
0.10188831 u
1
2
+
√
3
4
i 2.45366725
28 / 73
Harnack’s Principle
Using Harnack’s inequality we can can conclude that non-increasing
sequences of harmonic functions always converge. This is called Harnack’s
Theorem or Harnack’s Principle.
Theorem
Let u1 ≤ u2 ≤ . . . be harmonic functions on a connected open set U ⊆ C.
Then either uj → ∞ uniformly on compact sets or there is a harmonic
function u on U such that uj → u uniformly on compact sets.
Alternatively
Let u1 ≤ u2 ≤ . . . be harmonic functions on a connected open set U ⊆ C.
If for some point z0 ∈ U it happens that limn→∞ un(z0) = ∞, then
limn→∞ un(z) = ∞ uniformly on every compact set in U. However if
limn→∞ un(z0) = L ≤ ∞ then un(z) converges throughout U to a
harmonic function u uniformly on compact subsets of U.
29 / 73
Subharmonic Functions
We now know what a harmonic function is and some of its properties so
lets define what a subharmonic function is and look at some of its
properties.
30 / 73
Subharmonic Functions
We now know what a harmonic function is and some of its properties so
lets define what a subharmonic function is and look at some of its
properties.
Definition
Let U ⊆ C be an open set and f a real valued continuous function on U.
Suppose that for each D(P, r) ⊆ U and every real valued harmonic
function h defined on a neighborhood of D(P, r) which satisfies f ≤ h on
∂D(P, r) , it holds that f ≤ h on D(P, r). Then f is said to be
subharmonic on U.
31 / 73
Subharmonic Functions
Example
Is the function |z|2 − 1 subharmonic on D(0, 1)?
It is hard to decide this based on the definition of a subharmonic function
thus we have a proposition to help of decide this.
32 / 73
Subharmonic Functions
Example
Is the function |z|2 − 1 subharmonic on D(0, 1)?
It is hard to decide this based on the definition of a subharmonic function
thus we have a proposition to help of decide this.
Proposition
If f is C2 on an open set U, f is subharmonic if and only if ∆f ≥ 0 on U.
33 / 73
Subharmonic Functions
Example
Is the function f (z) = |z|2 − 1 subharmonic on D(0, 1)?
∂f
∂x
= 2x
∂2f
∂x2
= 2
∂f
∂y
= 2y
∂2f
∂y2
= 2
Therefore ∆u = ∂2
∂x2 + ∂2
∂y2 u = 2 + 2 = 4 > 0 and f is subharmonic.
34 / 73
Submean-value Property
Proposition
Let f : U → R be continuous. Suppose that, for each D(P, r) ⊆ U,
f (P) ≤
1
2π
2π
0
f (P + reiθ
)dθ. (∗)
Then f is subharmonic.
Conversely, if f : U → R is a (continuous) subharmonic function and if
D(P, r) ⊆ U, then the inequality (∗) holds.
35 / 73
Submean-value Property
36 / 73
Maximum Principle for Subharmonic Functions
Proposition
If f is subharmonic on U, and if there is a P ∈ U such that f (P) ≥ f (z)
for all z ∈ U, then f is constant.
37 / 73
Maximum Principle for Subharmonic Functions
Proposition
If f is subharmonic on U, and if there is a P ∈ U such that f (P) ≥ f (z)
for all z ∈ U, then f is constant.
Q: Is there a minimum principle for subharmonic functions?
38 / 73
Maximum Principle for Subharmonic Functions
Proposition
If f is subharmonic on U, and if there is a P ∈ U such that f (P) ≥ f (z)
for all z ∈ U, then f is constant.
Q: Is there a minimum principle for subharmonic functions?
A: No. By the the submean-value principle for subharmonic functions
there does not exist a minimum principle because the value of f can
be strictly smaller than the average on the corresoponding circle.
39 / 73
Barrier Functions
Now we will define a special function, that singles out a point P in the
boundary of a set, called a barrier.
40 / 73
Barrier Functions
Now we will define a special function, that singles out a point P in the
boundary of a set, called a barrier.
Definition
Let U ⊆ C be an open set and P ∈ ∂U. We call a function b : U → R a
barrier of U at P if
(I) b is continuous ;
(II) b is subharmonic on U;
(III) b ≤ 0;
(IV) {z ∈ ∂U : b(z) = 0} = {P}.
41 / 73
Barrier Function
Example
Let U = D(0, 1) and P = 1 + 0i.
42 / 73
Barrier Function
Example
Let U = D(0, 1) and P = 1 + 0i.
Then b(z) = x − 1 is a barrier for U at P, where x = Re z
43 / 73
Barrier Function
Example
Let U = D(0, 1) and P = 1 + 0i.
Then b(z) = x − 1 is a barrier for U at P, where x = Re z
In general if U ⊆ C is open and bounded. Let P ∈ U be the point in U
that is the furthest from 0.
44 / 73
Barrier Function
Example
Let U = D(0, 1) and P = 1 + 0i.
Then b(z) = x − 1 is a barrier for U at P, where x = Re z
In general if U ⊆ C is open and bounded. Let P ∈ U be the point in U
that is the furthest from 0.
Let r = |P| and θ0 = arg P, 0 ≤ θ0 ≤ 2π then
45 / 73
Barrier Function
Example
Let U = D(0, 1) and P = 1 + 0i.
Then b(z) = x − 1 is a barrier for U at P, where x = Re z
In general if U ⊆ C is open and bounded. Let P ∈ U be the point in U
that is the furthest from 0.
Let r = |P| and θ0 = arg P, 0 ≤ θ0 ≤ 2π then
z −→ Re(e−iθ0
z) − r
is a barrier for U at P.
46 / 73
Barrier Functions
47 / 73
Perr´on Method
Q: Does every set U ⊆ C with a given boundary condition have a
solution to the Dirichlet problem?
48 / 73
Perr´on Method
Q: Does every set U ⊆ C with a given boundary condition have a
solution to the Dirichlet problem?
A: No, for example
Example
Let U = D(0, 1)  {0}. Then ∂U = {z : |z| = 1} ∪ {0}. Set
f (z) =
1 if |z| = 1
0 if z = 0
49 / 73
Perr´on Method
Example
Let U = D(0, 1)  {0}. Then ∂U = {z : |z| = 1} ∪ {0}. Set
f (z) =
1 if |z| = 1
0 if z = 0.
f is continuous on ∂U, if there is a solution to the Dirichlet problem it
must be that u(z) = u(eiθz) for some fixed θ. This is because if u(z) is a
solution then u(eiθz) is a solution for a fixed θ and because the Dirichlet
problem as a unique solution then u(z) = u(eiθz).
50 / 73
Perr´on Method
Example (Cont.)
By using the polar form of the Laplacian, ∆ = 1
r
∂
∂r r ∂
∂r + 1
r2
∂2
∂θ2 , we see
that
0 = ∆u =
1
r
∂
∂r
r
∂u
∂r
(Independence of θ)
0 =
∂
∂r
r
∂u
∂r
(When r = 0)
Thus r ∂u
∂r = C for some constant C ∈ R.
51 / 73
Perr´on Method
Example (Cont.)
By using the polar form of the Laplacian, ∆ = 1
r
∂
∂r r ∂
∂r + 1
r2
∂2
∂θ2 , we see
that
0 = ∆u =
1
r
∂
∂r
r
∂u
∂r
(Independence of θ)
0 =
∂
∂r
r
∂u
∂r
(When r = 0)
Thus r ∂u
∂r = C for some constant C ∈ R. This is a simple ODE and thus
u = C log(r) + D for some constant D ∈ R.
52 / 73
Perr´on Method
Example (Cont.)
By using the polar form of the Laplacian, ∆ = 1
r
∂
∂r r ∂
∂r + 1
r2
∂2
∂θ2 , we see
that
0 = ∆u =
1
r
∂
∂r
r
∂u
∂r
(Independence of θ)
0 =
∂
∂r
r
∂u
∂r
(When r = 0)
Thus r ∂u
∂r = C for some constant C ∈ R. This is a simple ODE and thus
u = C log(r) + D for some constant D ∈ R.
There are no choices for C and D that will allow such a u to agree with
the given f on the boundary of U. Thus this Dirichlet problem cannot be
solved.
53 / 73
Perr´on Method
Q: Can we ever guarantee that a solution to the Dirichlet problem will
exist?
54 / 73
Perr´on Method
Q: Can we ever guarantee that a solution to the Dirichlet problem will
exist?
A: Yes!
Theorem
Let U be a bounded, connected open subset of C such that U has a
barrier bP for each P ∈ ∂U. Then the Dirichlet problem can always be
solved on U. That is, if f is continuous function on ∂U, then there is a
function u continuous on U, harmonic on U, such that u ∂U
= f . The
function u is uniquely determined by these conditions.
55 / 73
Perr´on Method
Q: So what went wrong this our example?
Example
Let U = D(0, 1)  {0}. Then ∂U = {z : |z| = 1} ∪ {0}. Set
f (z) =
1 if |z| = 1
0 if z = 0
56 / 73
Perr´on Method
Q: So what went wrong this our example?
Example
Let U = D(0, 1)  {0}. Then ∂U = {z : |z| = 1} ∪ {0}. Set
f (z) =
1 if |z| = 1
0 if z = 0
A: There does not exist a barrier function at the point z = 0 in the
boundary of U.
57 / 73
Perr´on Method
Theorem
Let U be a bounded, connected open subset of C such that U has a
barrier bP for each P ∈ ∂U. Then the Dirichlet problem can always be
solved on U. That is, if f is continuous function on ∂U, then there is a
function u continuous on U, harmonic on U, such that u ∂U
= f . The
function u is uniquely determined by these conditions.
58 / 73
Perr´on Method
Proof.
Without loss of generality assume that f is real valued. First set
S = {ψ : ψ is subharmonic of U and lim sup
U z→P
ψ(z) ≤ f (P), ∀P ∈ ∂U}
The set S is not empty because ∂U is compact and bounded below by
some real constant m and therefore ψ(z) ≡ m is in S
Define , for each z ∈ U
u(z) = sup
ψ∈S
ψ(z).
59 / 73
Perr´on Method
Proof.
Without loss of generality assume that f is real valued. First set
S = {ψ : ψ is subharmonic of U and lim sup
U z→P
ψ(z) ≤ f (P), ∀P ∈ ∂U}
The set S is not empty because ∂U is compact and bounded below by
some real constant m and therefore ψ(z) ≡ m is in S
Define , for each z ∈ U
u(z) = sup
ψ∈S
ψ(z).
I claim that u solves the Dirichlet problem for f and U. We can show this
in 3 steps.
60 / 73
Perr´on Method
Proof.
Without loss of generality assume that f is real valued. First set
S = {ψ : ψ is subharmonic of U and lim sup
U z→P
ψ(z) ≤ f (P), ∀P ∈ ∂U}
The set S is not empty because ∂U is compact and bounded below by
some real constant m and therefore ψ(z) ≡ m is in S
Define , for each z ∈ U
u(z) = sup
ψ∈S
ψ(z).
I claim that u solves the Dirichlet problem for f and U. We can show this
in 3 steps.
1 The function u is bounded above.
61 / 73
Perr´on Method
Proof.
Without loss of generality assume that f is real valued. First set
S = {ψ : ψ is subharmonic of U and lim sup
U z→P
ψ(z) ≤ f (P), ∀P ∈ ∂U}
The set S is not empty because ∂U is compact and bounded below by
some real constant m and therefore ψ(z) ≡ m is in S
Define , for each z ∈ U
u(z) = sup
ψ∈S
ψ(z).
I claim that u solves the Dirichlet problem for f and U. We can show this
in 3 steps.
1 The function u is bounded above.
2 The function u is harmonic.
62 / 73
Perr´on Method
Proof.
Without loss of generality assume that f is real valued. First set
S = {ψ : ψ is subharmonic of U and lim sup
U z→P
ψ(z) ≤ f (P), ∀P ∈ ∂U}
The set S is not empty because ∂U is compact and bounded below by
some real constant m and therefore ψ(z) ≡ m is in S
Define , for each z ∈ U
u(z) = sup
ψ∈S
ψ(z).
I claim that u solves the Dirichlet problem for f and U. We can show this
in 3 steps.
1 The function u is bounded above.
2 The function u is harmonic.
3 For each w ∈ ∂U, limU z→w u(z) = f (w).
63 / 73
Perr´on Method
Cont.
1 The function u is bounded above.
Let M = maxζ∈∂U f (ζ), take ψ ∈ S and ε > 0, and let
Eε = {z ∈ U : ψ(z) ≥ M + ε}. Assume that Eε is nonempty.
64 / 73
Perr´on Method
Cont.
1 The function u is bounded above.
Let M = maxζ∈∂U f (ζ), take ψ ∈ S and ε > 0, and let
Eε = {z ∈ U : ψ(z) ≥ M + ε}. Assume that Eε is nonempty.
It can be shown that Eε is closed.
65 / 73
Perr´on Method
Cont.
1 The function u is bounded above.
Let M = maxζ∈∂U f (ζ), take ψ ∈ S and ε > 0, and let
Eε = {z ∈ U : ψ(z) ≥ M + ε}. Assume that Eε is nonempty.
It can be shown that Eε is closed.
Since Eε ⊂ U and U is bounded then Eε is bounded then Eε is
compact. Since ψ is continuous and Eε is compact then then ψ
attains its maximum at some point P ∈ Eε. Since the max is at least
M + ε on Eε and ψ is less than M + ε outside Eε then ψ(P) is a max
for ψ on all of U. Therefore by the maximum principle ψ is constant.
Since P ∈ Eε then the constant must be greater than or equal to
M + ε which contradicts ψ ∈ S. So Eε must be empty and thus
ψ(ζ) ≤ M for all ζ ∈ U. Therefore u is bounded above.
66 / 73
Perr´on Method
Cont.
2 The function u is harmonic
Let D(P, r) ⊆ U and p ∈ D(P, r). By definition of u there exists a
sequence {ψj }∞
j=1 ⊆ S such that limj→∞ ψj (p) = u(p). Let
Ψn(z) = max{ψ1(z), . . . , ψn(z)}
for each z ∈ U. By properties of subharmonic functions Ψn(z) is
subharmonic on U and Ψ1 ≤ Ψ2 ≤ . . . . Let
Φn(z) =
Ψn(z) if z ∈ U  D(P, r)
Poisson integral of Ψn ∂D(P,r)
if z ∈ D(P, r)
By the submean-value property the function Φn is subharmonic on U
and also Φn ∈ S. Furthermore Φ1 ≤ Φ2 ≤ . . . on U by maximum
principle.
67 / 73
Perr´on Method
Cont.
Finally,
ψn(p) ≤ Ψn(p) ≤ Φn(p) ≤ u(p) (∗)
for each n and thus limn→∞ Φn(p) = u(p) By Harnack’s principle the
function {Φn} converse to a harmonic function Φ on the disk D(P, r).
Moreover Φ(P) = u(P) by (∗).
Take a point q ∈ D(P, r), there is a sequence {ρj }∞
j=1 ⊆ S such that
limj→∞ ρj (q) = u(q). Let ˜ρj (z) = max{ρj (z), ψj (z)} for z ∈ U and let
Λn(z) = max{˜ρ1(z), . . . , ˜ρn(z)}
for each z ∈ U. Again Λn is a subharmonic function on U with
Λ1 ≤ Λ2 ≤ . . . and Λn(q) → u(q).
68 / 73
Perr´on Method
Cont.
Define Hn(z) in the same way we defined Φn(z) before
Hn(z) =
Λn(z) if z ∈ U  D(P, r)
Poisson integral of Λn ∂D(P,r)
if z ∈ D(P, r)
{Hn} has a harmonic limit H on D(P, r) and H(q) = u(q). By design,
Φ(z) ≤ H(z) ≤ u(z) on D(P, r)
Hence H(p) = u(p) because Φ(p) = u(p). The harmonic function Φ − H
is such that Φ − H ≤ 0 on ∂D(P, r) and (Φ − H)(p) = 0. The maximum
principle says then that Φ ≡ H. Thus for any q ∈ D(P, r) the function H
agrees with the function Φ on p. Thus u(p) agrees with u(q) for each
point in D(P, r). Since P and r are chosen at random then we can
conclude that u is harmonic.
69 / 73
Perr´on Method
Cont.
3 For each w ∈ ∂U, limU z→w u(z) = f (w).
Fix w ∈ ∂U. For ε, k > 0 we have that g(z) = f (w) − ε + kbw (z) is
subharmonic for z ∈ U. Choose δ > 0, a function of ε, so that
f (z) > f (w) − ε whenever z ∈ ∂U ∩ D(w, δ). Thus
g(z) ≤ f (z) for z ∈ ∂U ∩ D(w, δ)
Note that bp(z) has a strictly negative upper bound on ∂U  D(w, δ).
Choose k, a function of ε, large enough such that
g(z) ≤ f (z) for z ∈ ∂U  D(w, δ)
Thus g ∈ S, by the maximum principle for subharmonic functions, so
g ≤ u and
f (w) − ε = lim
U z→w
g(z) ≤ lim inf
U z→w
u(z)
70 / 73
Perr´on Method
Cont.
By a similar argument lim supU z→w u(z) ≤ f (w) + ε. Thus
f (w) − ε ≤ lim inf
U z→w
u(z) ≤ lim sup
U z→w
u(z) ≤ f (w) + ε
Therefore we have the desired result limU z→w u(z) = f (w) : for w ∈ ∂U.
The final thing to show is that u is unique. This follows from the
maximum principle which says that the maximum for a harmonic function
occurs on the boundary.
71 / 73
References
Greene, Robert E., and Krantz, Steven G. Function Theory of One Complex
Variable. 3rd ed. 40. American Mathematical Society, 2006. Print.
72 / 73
Thank You
Thank you all for coming. I would
like to give a special thank you to
Dr. Edoh Amiran
73 / 73

More Related Content

What's hot

Partial differentiation
Partial differentiationPartial differentiation
Partial differentiationTanuj Parikh
 
Partial Differential Equation - Notes
Partial Differential Equation - NotesPartial Differential Equation - Notes
Partial Differential Equation - NotesDr. Nirav Vyas
 
Geometric and viscosity solutions for the Cauchy problem of first order
Geometric and viscosity solutions for the Cauchy problem of first orderGeometric and viscosity solutions for the Cauchy problem of first order
Geometric and viscosity solutions for the Cauchy problem of first orderJuliho Castillo
 
Partial Differential Equations, 3 simple examples
Partial Differential Equations, 3 simple examplesPartial Differential Equations, 3 simple examples
Partial Differential Equations, 3 simple examplesEnrique Valderrama
 
How to Solve a Partial Differential Equation on a surface
How to Solve a Partial Differential Equation on a surfaceHow to Solve a Partial Differential Equation on a surface
How to Solve a Partial Differential Equation on a surfacetr1987
 
Solving High-order Non-linear Partial Differential Equations by Modified q-Ho...
Solving High-order Non-linear Partial Differential Equations by Modified q-Ho...Solving High-order Non-linear Partial Differential Equations by Modified q-Ho...
Solving High-order Non-linear Partial Differential Equations by Modified q-Ho...BRNSS Publication Hub
 
Partial Differentiation & Application
Partial Differentiation & Application Partial Differentiation & Application
Partial Differentiation & Application Yana Qlah
 
Physical Chemistry Homework Help
Physical Chemistry Homework HelpPhysical Chemistry Homework Help
Physical Chemistry Homework HelpEdu Assignment Help
 
Fixed points of contractive and Geraghty contraction mappings under the influ...
Fixed points of contractive and Geraghty contraction mappings under the influ...Fixed points of contractive and Geraghty contraction mappings under the influ...
Fixed points of contractive and Geraghty contraction mappings under the influ...IJERA Editor
 
Advanced Microeconomics - Lecture Slides
Advanced Microeconomics - Lecture SlidesAdvanced Microeconomics - Lecture Slides
Advanced Microeconomics - Lecture SlidesYosuke YASUDA
 
International journal of engineering and mathematical modelling vol2 no3_2015_2
International journal of engineering and mathematical modelling vol2 no3_2015_2International journal of engineering and mathematical modelling vol2 no3_2015_2
International journal of engineering and mathematical modelling vol2 no3_2015_2IJEMM
 
Applications of differential equations by shahzad
Applications of differential equations by shahzadApplications of differential equations by shahzad
Applications of differential equations by shahzadbiotech energy pvt limited
 
Introduction to Decision Making Theory
Introduction to Decision Making TheoryIntroduction to Decision Making Theory
Introduction to Decision Making TheoryYosuke YASUDA
 
Simultaneous differential equations
Simultaneous differential equationsSimultaneous differential equations
Simultaneous differential equationsShubhi Jain
 
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...Avichai Cohen
 

What's hot (20)

Partial differentiation
Partial differentiationPartial differentiation
Partial differentiation
 
Partial Differential Equation - Notes
Partial Differential Equation - NotesPartial Differential Equation - Notes
Partial Differential Equation - Notes
 
Geometric and viscosity solutions for the Cauchy problem of first order
Geometric and viscosity solutions for the Cauchy problem of first orderGeometric and viscosity solutions for the Cauchy problem of first order
Geometric and viscosity solutions for the Cauchy problem of first order
 
Partial Differential Equations, 3 simple examples
Partial Differential Equations, 3 simple examplesPartial Differential Equations, 3 simple examples
Partial Differential Equations, 3 simple examples
 
Legendre functions
Legendre functionsLegendre functions
Legendre functions
 
How to Solve a Partial Differential Equation on a surface
How to Solve a Partial Differential Equation on a surfaceHow to Solve a Partial Differential Equation on a surface
How to Solve a Partial Differential Equation on a surface
 
Solving High-order Non-linear Partial Differential Equations by Modified q-Ho...
Solving High-order Non-linear Partial Differential Equations by Modified q-Ho...Solving High-order Non-linear Partial Differential Equations by Modified q-Ho...
Solving High-order Non-linear Partial Differential Equations by Modified q-Ho...
 
Thesis
ThesisThesis
Thesis
 
Reduction forumla
Reduction forumlaReduction forumla
Reduction forumla
 
Partial
Partial Partial
Partial
 
Partial Differentiation & Application
Partial Differentiation & Application Partial Differentiation & Application
Partial Differentiation & Application
 
Physical Chemistry Homework Help
Physical Chemistry Homework HelpPhysical Chemistry Homework Help
Physical Chemistry Homework Help
 
Fixed points of contractive and Geraghty contraction mappings under the influ...
Fixed points of contractive and Geraghty contraction mappings under the influ...Fixed points of contractive and Geraghty contraction mappings under the influ...
Fixed points of contractive and Geraghty contraction mappings under the influ...
 
Advanced Microeconomics - Lecture Slides
Advanced Microeconomics - Lecture SlidesAdvanced Microeconomics - Lecture Slides
Advanced Microeconomics - Lecture Slides
 
International journal of engineering and mathematical modelling vol2 no3_2015_2
International journal of engineering and mathematical modelling vol2 no3_2015_2International journal of engineering and mathematical modelling vol2 no3_2015_2
International journal of engineering and mathematical modelling vol2 no3_2015_2
 
Slides mc gill-v4
Slides mc gill-v4Slides mc gill-v4
Slides mc gill-v4
 
Applications of differential equations by shahzad
Applications of differential equations by shahzadApplications of differential equations by shahzad
Applications of differential equations by shahzad
 
Introduction to Decision Making Theory
Introduction to Decision Making TheoryIntroduction to Decision Making Theory
Introduction to Decision Making Theory
 
Simultaneous differential equations
Simultaneous differential equationsSimultaneous differential equations
Simultaneous differential equations
 
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
Amirim Project - Threshold Functions in Random Simplicial Complexes - Avichai...
 

Similar to Solving the Dirichlet Problem Using Subharmonic Functions

590-Article Text.pdf
590-Article Text.pdf590-Article Text.pdf
590-Article Text.pdfBenoitValea
 
590-Article Text.pdf
590-Article Text.pdf590-Article Text.pdf
590-Article Text.pdfBenoitValea
 
Natalini nse slide_giu2013
Natalini nse slide_giu2013Natalini nse slide_giu2013
Natalini nse slide_giu2013Madd Maths
 
Hyperfunction method for numerical integration and Fredholm integral equation...
Hyperfunction method for numerical integration and Fredholm integral equation...Hyperfunction method for numerical integration and Fredholm integral equation...
Hyperfunction method for numerical integration and Fredholm integral equation...HidenoriOgata
 
Density theorems for Euclidean point configurations
Density theorems for Euclidean point configurationsDensity theorems for Euclidean point configurations
Density theorems for Euclidean point configurationsVjekoslavKovac1
 
IVR - Chapter 1 - Introduction
IVR - Chapter 1 - IntroductionIVR - Chapter 1 - Introduction
IVR - Chapter 1 - IntroductionCharles Deledalle
 
Partial diferential good
Partial diferential goodPartial diferential good
Partial diferential goodgenntmbr
 
Numerical integration based on the hyperfunction theory
Numerical integration based on the hyperfunction theoryNumerical integration based on the hyperfunction theory
Numerical integration based on the hyperfunction theoryHidenoriOgata
 
Digital Text Book :POTENTIAL THEORY AND ELLIPTIC PARTIAL DIFFERENTIAL EQUATIONS
Digital Text Book :POTENTIAL THEORY AND ELLIPTIC PARTIAL DIFFERENTIAL EQUATIONS Digital Text Book :POTENTIAL THEORY AND ELLIPTIC PARTIAL DIFFERENTIAL EQUATIONS
Digital Text Book :POTENTIAL THEORY AND ELLIPTIC PARTIAL DIFFERENTIAL EQUATIONS Baasilroy
 
1 hofstad
1 hofstad1 hofstad
1 hofstadYandex
 
Maximum likelihood estimation of regularisation parameters in inverse problem...
Maximum likelihood estimation of regularisation parameters in inverse problem...Maximum likelihood estimation of regularisation parameters in inverse problem...
Maximum likelihood estimation of regularisation parameters in inverse problem...Valentin De Bortoli
 
Litvinenko_RWTH_UQ_Seminar_talk.pdf
Litvinenko_RWTH_UQ_Seminar_talk.pdfLitvinenko_RWTH_UQ_Seminar_talk.pdf
Litvinenko_RWTH_UQ_Seminar_talk.pdfAlexander Litvinenko
 
Crib Sheet AP Calculus AB and BC exams
Crib Sheet AP Calculus AB and BC examsCrib Sheet AP Calculus AB and BC exams
Crib Sheet AP Calculus AB and BC examsA Jorge Garcia
 
OrthogonalFunctionsPaper
OrthogonalFunctionsPaperOrthogonalFunctionsPaper
OrthogonalFunctionsPaperTyler Otto
 
cmftJYeZhuanTalk.pdf
cmftJYeZhuanTalk.pdfcmftJYeZhuanTalk.pdf
cmftJYeZhuanTalk.pdfjyjyzr69t7
 
Backpropagation in Convolutional Neural Network
Backpropagation in Convolutional Neural NetworkBackpropagation in Convolutional Neural Network
Backpropagation in Convolutional Neural NetworkHiroshi Kuwajima
 
Functional analysis in mechanics 2e
Functional analysis in mechanics  2eFunctional analysis in mechanics  2e
Functional analysis in mechanics 2eSpringer
 
Functional analysis in mechanics
Functional analysis in mechanicsFunctional analysis in mechanics
Functional analysis in mechanicsSpringer
 

Similar to Solving the Dirichlet Problem Using Subharmonic Functions (20)

590-Article Text.pdf
590-Article Text.pdf590-Article Text.pdf
590-Article Text.pdf
 
590-Article Text.pdf
590-Article Text.pdf590-Article Text.pdf
590-Article Text.pdf
 
Natalini nse slide_giu2013
Natalini nse slide_giu2013Natalini nse slide_giu2013
Natalini nse slide_giu2013
 
Hyperfunction method for numerical integration and Fredholm integral equation...
Hyperfunction method for numerical integration and Fredholm integral equation...Hyperfunction method for numerical integration and Fredholm integral equation...
Hyperfunction method for numerical integration and Fredholm integral equation...
 
Density theorems for Euclidean point configurations
Density theorems for Euclidean point configurationsDensity theorems for Euclidean point configurations
Density theorems for Euclidean point configurations
 
IVR - Chapter 1 - Introduction
IVR - Chapter 1 - IntroductionIVR - Chapter 1 - Introduction
IVR - Chapter 1 - Introduction
 
Partial diferential good
Partial diferential goodPartial diferential good
Partial diferential good
 
Numerical integration based on the hyperfunction theory
Numerical integration based on the hyperfunction theoryNumerical integration based on the hyperfunction theory
Numerical integration based on the hyperfunction theory
 
Digital Text Book :POTENTIAL THEORY AND ELLIPTIC PARTIAL DIFFERENTIAL EQUATIONS
Digital Text Book :POTENTIAL THEORY AND ELLIPTIC PARTIAL DIFFERENTIAL EQUATIONS Digital Text Book :POTENTIAL THEORY AND ELLIPTIC PARTIAL DIFFERENTIAL EQUATIONS
Digital Text Book :POTENTIAL THEORY AND ELLIPTIC PARTIAL DIFFERENTIAL EQUATIONS
 
1 hofstad
1 hofstad1 hofstad
1 hofstad
 
Maximum likelihood estimation of regularisation parameters in inverse problem...
Maximum likelihood estimation of regularisation parameters in inverse problem...Maximum likelihood estimation of regularisation parameters in inverse problem...
Maximum likelihood estimation of regularisation parameters in inverse problem...
 
Litvinenko_RWTH_UQ_Seminar_talk.pdf
Litvinenko_RWTH_UQ_Seminar_talk.pdfLitvinenko_RWTH_UQ_Seminar_talk.pdf
Litvinenko_RWTH_UQ_Seminar_talk.pdf
 
VECTOR CALCULUS
VECTOR CALCULUS VECTOR CALCULUS
VECTOR CALCULUS
 
Ece3075 a 8
Ece3075 a 8Ece3075 a 8
Ece3075 a 8
 
Crib Sheet AP Calculus AB and BC exams
Crib Sheet AP Calculus AB and BC examsCrib Sheet AP Calculus AB and BC exams
Crib Sheet AP Calculus AB and BC exams
 
OrthogonalFunctionsPaper
OrthogonalFunctionsPaperOrthogonalFunctionsPaper
OrthogonalFunctionsPaper
 
cmftJYeZhuanTalk.pdf
cmftJYeZhuanTalk.pdfcmftJYeZhuanTalk.pdf
cmftJYeZhuanTalk.pdf
 
Backpropagation in Convolutional Neural Network
Backpropagation in Convolutional Neural NetworkBackpropagation in Convolutional Neural Network
Backpropagation in Convolutional Neural Network
 
Functional analysis in mechanics 2e
Functional analysis in mechanics  2eFunctional analysis in mechanics  2e
Functional analysis in mechanics 2e
 
Functional analysis in mechanics
Functional analysis in mechanicsFunctional analysis in mechanics
Functional analysis in mechanics
 

Solving the Dirichlet Problem Using Subharmonic Functions

  • 1. Subharmonic Functions and the Solution to the Dirichlet Problem Author: Bryan R. Clark Advisor: Dr. Edoh Amiran Western Washington University brclark9@gmail.com Thursday 26th September, 2013 1 / 73
  • 2. Overview 1 Harmonic Functions Properties of Harmonic Functions Dirichlet Problem for the Disk Harnack’s Principle 2 Subharmonic Functions Properties Barrier Function Perr´on Method 2 / 73
  • 3. Harmonic Functions Definition A real-valued function u : U → R on an open set U ⊆ C is harmonic if it is C2 on U and ∆u ≡ 0 there where the Laplacian ∆u is defined by ∆u = ∂2 ∂x2 + ∂2 ∂y2 u. 3 / 73
  • 4. Harmonic Functions Example Is u(x, y) = e3x cos(3y) harmonic? 4 / 73
  • 5. Harmonic Functions Example Is u(x, y) = e3x cos(3y) harmonic? ∂u ∂x = 3e3x cos(3y) ∂2u ∂x2 = 9e3x cos(3y) ∂u ∂y = −3e3x sin(3y) ∂2u ∂y2 = −9e3x cos(3y) Therefore ∆u = ∂2 ∂x2 + ∂2 ∂y2 u = 9e3x cos(3y) − 9e3x cos(3y) = 0 thus u is harmonic. 5 / 73
  • 6. Maximum/Minimum Principle Theorem If u : U → R is harmonic on a connected open set U and if there is a point P0 ∈ U with the property that u(P0) = supQ∈U u(Q) or u(P0) = infQ∈U u(Q), then u is constant on U. 6 / 73
  • 7. Maximum/Minimum Principle Theorem If u : U → R is harmonic on a connected open set U and if there is a point P0 ∈ U with the property that u(P0) = supQ∈U u(Q) or u(P0) = infQ∈U u(Q), then u is constant on U. An immediate consequence of this theorem is the following corollary. Corollary Let U ⊆ C be a bounded, connected open set and let u be a continuous, real valued function on the closure U of U that is harmonic on U. Then max u U = max u ∂U min u U = min u ∂U . 7 / 73
  • 8. Maximum/Minimum Principle Lets prove this corollary. Proof. A continuous function on a closed bounded set attains its maximum. Thus there exists a P ∈ U such that u(P) is maximal. If P ∈ U then by the maximum principle u is constant and thus attains its maximum. If P ∈ ∂U then the result is certainly true. 8 / 73
  • 9. Mean-value Property Theorem Suppose that u : U → R is a harmonic function on an open set U ⊆ C and that D(P, r) ⊆ U for some r > 0 then u(P) = 1 2π 2π 0 u(P + reiθ )dθ. 9 / 73
  • 10. Mean-value Property Theorem Suppose that u : U → R is a harmonic function on an open set U ⊆ C and that D(P, r) ⊆ U for some r > 0 then u(P) = 1 2π 2π 0 u(P + reiθ )dθ. Q: What does this mean? 10 / 73
  • 12. Mean-value Property A: Q: What does this remind you of? 12 / 73
  • 13. Mean-value Property A: Q: What does this remind you of? A: The Cauchy integral formula. 13 / 73
  • 14. Mean-value Property Example Let u(x, y) = x2 − y2, harmonic, and the disk D(P, 1) where P = 1 + i. Then u(1, 1) = 1 2π 2π 0 u(1 + cos(θ), 1 + sin(θ)) dθ = 1 2π 2π 0 [1 + cos(θ)]2 − [1 + sin(θ)]2 dθ = 1 2π 2π 0 2 cos(θ) − 2 sin(θ) + cos(2θ) dθ = 1 2π 2π 0 2 cos(θ) dθ − 1 2π 2π 0 2 sin(θ) dθ + 1 2π 2π 0 cos(2θ) dθ = 0 14 / 73
  • 15. Poisson Integral Formula Theorem Let u : U → R be a harmonic function on a neighborhood of D(0, 1). Then, for any point a ∈ D(0, 1), u(a) = 1 2π 2π 0 u(eiψ ) 1 − |a|2 |a − eiψ|2 dψ. 15 / 73
  • 16. Poisson Integral Formula Theorem Let u : U → R be a harmonic function on a neighborhood of D(0, 1). Then, for any point a ∈ D(0, 1), u(a) = 1 2π 2π 0 u(eiψ ) 1 − |a|2 |a − eiψ|2 dψ. The Poisson integral formula says that the value u(reiψ) for some interior point reiψ is a weighted average of the values of u on the boundary of the circle where the weight P(θ − ψ) = 1 2π · 1 − r2 1 + r2 − 2r cos(θ − ψ) > 0 and has total weight 1. 16 / 73
  • 17. Dirichlet Problem for the Disk Definition Let D denote the unit disk. Given a bounded, piecewise continuous function on the boundary of the disk, u : ∂D → R is there a harmonic function on the disk, u : D → R that extends to match the original u on the boundary of the disk, the extension begin continuous other than the original discontinuities? 17 / 73
  • 18. Dirichlet problem for the Disk Does there always exist a solution to the Dirichlet problem on the disk? Yes or No 18 / 73
  • 19. Solution to the Dirichlet Problem for the Disk Theorem Let f be a continuous function on ∂D(0, 1). Define u(z) =    1 2π 2π 0 f (eiψ) · 1−|z|2 |z−eiψ|2 dψ if z ∈ D(0, 1) f (z) if z ∈ ∂D(0, 1) Then u is continuous on D(0, 1) and harmonic on D(0, 1). 19 / 73
  • 20. Solution to the Dirichlet Problem for the Disk Proof. It can be shown that 1 − |z|2 |z − eiθ|2 = eiθ eiθ − z + e−iθ e−iθ − ¯z − 1. When z ∈ D(0, 1), u(z) = 1 2π 2π 0 f (eiθ ) eiθ eiθ − z dθ + 1 2π 2π 0 f (eiθ ) e−iθ e−iθ − ¯z dθ − 1 2π 2π 0 f (eiθ )dθ 20 / 73
  • 21. Solution to the Dirichlet Problem for the Disk Continued. Using the fact that ∆ = 4 ∂2 ∂z∂¯z we conclude that the first integral is harmonic. ∂ ∂¯z 1 2π 2π 0 f (eiθ ) eiθ eiθ − z dθ = 1 2π 2π 0 f (eiθ ) ∂ ∂¯z eiθ eiθ − z dθ = 0 21 / 73
  • 22. Solution to the Dirichlet Problem for the Disk Continued. Using the fact that ∆ = 4 ∂2 ∂z∂¯z we conclude that the first integral is harmonic. ∂ ∂¯z 1 2π 2π 0 f (eiθ ) eiθ eiθ − z dθ = 1 2π 2π 0 f (eiθ ) ∂ ∂¯z eiθ eiθ − z dθ = 0 Since eiθ eiθ−z is holomorphic. Thus it satisfies the criteria for being harmonic. The second and third integrals are harmonic by the same argument. So we have a sum of harmonic functions which is harmonic. Therefore if z ∈ D(0, 1) then u is harmonic. 22 / 73
  • 23. Solution to the Dirichlet Problem for the Disk Continued. Now I will convince you that the function u(z) is continuous. When z ∈ D(0, 1) we want to show that u is continuous. The Poisson integral formula says that the value u(reiθ) for some interior point reiθ is a weighted average of the values of f on the boundary of the circle where the weight P(θ − ψ) > 0 and has total weight 1. We see that if we fix θ and take the limit limr→1− Pr (θ − ψ) we get 0 given θ = ψ. If r is very close to 1, i.e. eiθ very close to the boundary, then Pr (θ − ψ) is close to 0 except when ψ is close to θ. We are taking a weighted average with most of the weight concentrated near the point eiθ. Thus it makes sense that u(reiθ) converges to f (eiθ). 23 / 73
  • 24. Solution to the Dirichlet Problem for the Disk Example Find the function u harmonic on the unit disk, D(0, 1) = {z : |z| < 1} and takes on the boundary vales f (t) =    π for π 2 < t < π 0 for − π 2 < t < π 2 −π for − π < t < −π 2 24 / 73
  • 25. Solution to the Dirichlet Problem for the Disk Example (Continued) Solution: u(z) = − tan−1 1+r 1−r tan θ+π 2 + tan−1 1+r 1−r tan 2θ+π 4 π + tan−1 1+r 1−r tan 2θ−π 4 − tan−1 1+r 1−r tan θ−π 2 π if z ∈ D(0, 1), and u(z) = f (z) if z ∈ ∂D(0, 1) 25 / 73
  • 26. The Harnack Inequality Proposition Let u be a nonnegative, harmonic function on a neighborhood of D(P, R). Then for any z ∈ D(P, R), R − |z − P| R + |z − p| · u(P) ≤ u(z) ≤ R + |z − P| R − |z − P| · u(P). 26 / 73
  • 27. The Harnack Inequality Proposition Let u be a nonnegative, harmonic function on a neighborhood of D(P, R). Then for any z ∈ D(P, R), R − |z − P| R + |z − p| · u(P) ≤ u(z) ≤ R + |z − P| R − |z − P| · u(P). Harnack’s inequality is an inequality relating the values of a positive harmonic function at two points. 27 / 73
  • 28. The Harnack Inequality Example Let u(x, y) = 1 2(x + 1) and which is nonnegative on D(0, 1). Our upper and lower bounds for a point z = 1 2 + √ 3 4 i are, R − |z| R + |z| · u(0) ≤ u(z) ≤ R + |z| R − |z| · u(0). 1 − |1 2 + √ 3 4 | 1 + |1 2 + √ 3 4 | · 0.5 u 1 2 + √ 3 4 i 1 + |1 2 + √ 3 4 | 1 − |1 2 + √ 3 4 | · 0.5. 0.10188831 u 1 2 + √ 3 4 i 2.45366725 28 / 73
  • 29. Harnack’s Principle Using Harnack’s inequality we can can conclude that non-increasing sequences of harmonic functions always converge. This is called Harnack’s Theorem or Harnack’s Principle. Theorem Let u1 ≤ u2 ≤ . . . be harmonic functions on a connected open set U ⊆ C. Then either uj → ∞ uniformly on compact sets or there is a harmonic function u on U such that uj → u uniformly on compact sets. Alternatively Let u1 ≤ u2 ≤ . . . be harmonic functions on a connected open set U ⊆ C. If for some point z0 ∈ U it happens that limn→∞ un(z0) = ∞, then limn→∞ un(z) = ∞ uniformly on every compact set in U. However if limn→∞ un(z0) = L ≤ ∞ then un(z) converges throughout U to a harmonic function u uniformly on compact subsets of U. 29 / 73
  • 30. Subharmonic Functions We now know what a harmonic function is and some of its properties so lets define what a subharmonic function is and look at some of its properties. 30 / 73
  • 31. Subharmonic Functions We now know what a harmonic function is and some of its properties so lets define what a subharmonic function is and look at some of its properties. Definition Let U ⊆ C be an open set and f a real valued continuous function on U. Suppose that for each D(P, r) ⊆ U and every real valued harmonic function h defined on a neighborhood of D(P, r) which satisfies f ≤ h on ∂D(P, r) , it holds that f ≤ h on D(P, r). Then f is said to be subharmonic on U. 31 / 73
  • 32. Subharmonic Functions Example Is the function |z|2 − 1 subharmonic on D(0, 1)? It is hard to decide this based on the definition of a subharmonic function thus we have a proposition to help of decide this. 32 / 73
  • 33. Subharmonic Functions Example Is the function |z|2 − 1 subharmonic on D(0, 1)? It is hard to decide this based on the definition of a subharmonic function thus we have a proposition to help of decide this. Proposition If f is C2 on an open set U, f is subharmonic if and only if ∆f ≥ 0 on U. 33 / 73
  • 34. Subharmonic Functions Example Is the function f (z) = |z|2 − 1 subharmonic on D(0, 1)? ∂f ∂x = 2x ∂2f ∂x2 = 2 ∂f ∂y = 2y ∂2f ∂y2 = 2 Therefore ∆u = ∂2 ∂x2 + ∂2 ∂y2 u = 2 + 2 = 4 > 0 and f is subharmonic. 34 / 73
  • 35. Submean-value Property Proposition Let f : U → R be continuous. Suppose that, for each D(P, r) ⊆ U, f (P) ≤ 1 2π 2π 0 f (P + reiθ )dθ. (∗) Then f is subharmonic. Conversely, if f : U → R is a (continuous) subharmonic function and if D(P, r) ⊆ U, then the inequality (∗) holds. 35 / 73
  • 37. Maximum Principle for Subharmonic Functions Proposition If f is subharmonic on U, and if there is a P ∈ U such that f (P) ≥ f (z) for all z ∈ U, then f is constant. 37 / 73
  • 38. Maximum Principle for Subharmonic Functions Proposition If f is subharmonic on U, and if there is a P ∈ U such that f (P) ≥ f (z) for all z ∈ U, then f is constant. Q: Is there a minimum principle for subharmonic functions? 38 / 73
  • 39. Maximum Principle for Subharmonic Functions Proposition If f is subharmonic on U, and if there is a P ∈ U such that f (P) ≥ f (z) for all z ∈ U, then f is constant. Q: Is there a minimum principle for subharmonic functions? A: No. By the the submean-value principle for subharmonic functions there does not exist a minimum principle because the value of f can be strictly smaller than the average on the corresoponding circle. 39 / 73
  • 40. Barrier Functions Now we will define a special function, that singles out a point P in the boundary of a set, called a barrier. 40 / 73
  • 41. Barrier Functions Now we will define a special function, that singles out a point P in the boundary of a set, called a barrier. Definition Let U ⊆ C be an open set and P ∈ ∂U. We call a function b : U → R a barrier of U at P if (I) b is continuous ; (II) b is subharmonic on U; (III) b ≤ 0; (IV) {z ∈ ∂U : b(z) = 0} = {P}. 41 / 73
  • 42. Barrier Function Example Let U = D(0, 1) and P = 1 + 0i. 42 / 73
  • 43. Barrier Function Example Let U = D(0, 1) and P = 1 + 0i. Then b(z) = x − 1 is a barrier for U at P, where x = Re z 43 / 73
  • 44. Barrier Function Example Let U = D(0, 1) and P = 1 + 0i. Then b(z) = x − 1 is a barrier for U at P, where x = Re z In general if U ⊆ C is open and bounded. Let P ∈ U be the point in U that is the furthest from 0. 44 / 73
  • 45. Barrier Function Example Let U = D(0, 1) and P = 1 + 0i. Then b(z) = x − 1 is a barrier for U at P, where x = Re z In general if U ⊆ C is open and bounded. Let P ∈ U be the point in U that is the furthest from 0. Let r = |P| and θ0 = arg P, 0 ≤ θ0 ≤ 2π then 45 / 73
  • 46. Barrier Function Example Let U = D(0, 1) and P = 1 + 0i. Then b(z) = x − 1 is a barrier for U at P, where x = Re z In general if U ⊆ C is open and bounded. Let P ∈ U be the point in U that is the furthest from 0. Let r = |P| and θ0 = arg P, 0 ≤ θ0 ≤ 2π then z −→ Re(e−iθ0 z) − r is a barrier for U at P. 46 / 73
  • 48. Perr´on Method Q: Does every set U ⊆ C with a given boundary condition have a solution to the Dirichlet problem? 48 / 73
  • 49. Perr´on Method Q: Does every set U ⊆ C with a given boundary condition have a solution to the Dirichlet problem? A: No, for example Example Let U = D(0, 1) {0}. Then ∂U = {z : |z| = 1} ∪ {0}. Set f (z) = 1 if |z| = 1 0 if z = 0 49 / 73
  • 50. Perr´on Method Example Let U = D(0, 1) {0}. Then ∂U = {z : |z| = 1} ∪ {0}. Set f (z) = 1 if |z| = 1 0 if z = 0. f is continuous on ∂U, if there is a solution to the Dirichlet problem it must be that u(z) = u(eiθz) for some fixed θ. This is because if u(z) is a solution then u(eiθz) is a solution for a fixed θ and because the Dirichlet problem as a unique solution then u(z) = u(eiθz). 50 / 73
  • 51. Perr´on Method Example (Cont.) By using the polar form of the Laplacian, ∆ = 1 r ∂ ∂r r ∂ ∂r + 1 r2 ∂2 ∂θ2 , we see that 0 = ∆u = 1 r ∂ ∂r r ∂u ∂r (Independence of θ) 0 = ∂ ∂r r ∂u ∂r (When r = 0) Thus r ∂u ∂r = C for some constant C ∈ R. 51 / 73
  • 52. Perr´on Method Example (Cont.) By using the polar form of the Laplacian, ∆ = 1 r ∂ ∂r r ∂ ∂r + 1 r2 ∂2 ∂θ2 , we see that 0 = ∆u = 1 r ∂ ∂r r ∂u ∂r (Independence of θ) 0 = ∂ ∂r r ∂u ∂r (When r = 0) Thus r ∂u ∂r = C for some constant C ∈ R. This is a simple ODE and thus u = C log(r) + D for some constant D ∈ R. 52 / 73
  • 53. Perr´on Method Example (Cont.) By using the polar form of the Laplacian, ∆ = 1 r ∂ ∂r r ∂ ∂r + 1 r2 ∂2 ∂θ2 , we see that 0 = ∆u = 1 r ∂ ∂r r ∂u ∂r (Independence of θ) 0 = ∂ ∂r r ∂u ∂r (When r = 0) Thus r ∂u ∂r = C for some constant C ∈ R. This is a simple ODE and thus u = C log(r) + D for some constant D ∈ R. There are no choices for C and D that will allow such a u to agree with the given f on the boundary of U. Thus this Dirichlet problem cannot be solved. 53 / 73
  • 54. Perr´on Method Q: Can we ever guarantee that a solution to the Dirichlet problem will exist? 54 / 73
  • 55. Perr´on Method Q: Can we ever guarantee that a solution to the Dirichlet problem will exist? A: Yes! Theorem Let U be a bounded, connected open subset of C such that U has a barrier bP for each P ∈ ∂U. Then the Dirichlet problem can always be solved on U. That is, if f is continuous function on ∂U, then there is a function u continuous on U, harmonic on U, such that u ∂U = f . The function u is uniquely determined by these conditions. 55 / 73
  • 56. Perr´on Method Q: So what went wrong this our example? Example Let U = D(0, 1) {0}. Then ∂U = {z : |z| = 1} ∪ {0}. Set f (z) = 1 if |z| = 1 0 if z = 0 56 / 73
  • 57. Perr´on Method Q: So what went wrong this our example? Example Let U = D(0, 1) {0}. Then ∂U = {z : |z| = 1} ∪ {0}. Set f (z) = 1 if |z| = 1 0 if z = 0 A: There does not exist a barrier function at the point z = 0 in the boundary of U. 57 / 73
  • 58. Perr´on Method Theorem Let U be a bounded, connected open subset of C such that U has a barrier bP for each P ∈ ∂U. Then the Dirichlet problem can always be solved on U. That is, if f is continuous function on ∂U, then there is a function u continuous on U, harmonic on U, such that u ∂U = f . The function u is uniquely determined by these conditions. 58 / 73
  • 59. Perr´on Method Proof. Without loss of generality assume that f is real valued. First set S = {ψ : ψ is subharmonic of U and lim sup U z→P ψ(z) ≤ f (P), ∀P ∈ ∂U} The set S is not empty because ∂U is compact and bounded below by some real constant m and therefore ψ(z) ≡ m is in S Define , for each z ∈ U u(z) = sup ψ∈S ψ(z). 59 / 73
  • 60. Perr´on Method Proof. Without loss of generality assume that f is real valued. First set S = {ψ : ψ is subharmonic of U and lim sup U z→P ψ(z) ≤ f (P), ∀P ∈ ∂U} The set S is not empty because ∂U is compact and bounded below by some real constant m and therefore ψ(z) ≡ m is in S Define , for each z ∈ U u(z) = sup ψ∈S ψ(z). I claim that u solves the Dirichlet problem for f and U. We can show this in 3 steps. 60 / 73
  • 61. Perr´on Method Proof. Without loss of generality assume that f is real valued. First set S = {ψ : ψ is subharmonic of U and lim sup U z→P ψ(z) ≤ f (P), ∀P ∈ ∂U} The set S is not empty because ∂U is compact and bounded below by some real constant m and therefore ψ(z) ≡ m is in S Define , for each z ∈ U u(z) = sup ψ∈S ψ(z). I claim that u solves the Dirichlet problem for f and U. We can show this in 3 steps. 1 The function u is bounded above. 61 / 73
  • 62. Perr´on Method Proof. Without loss of generality assume that f is real valued. First set S = {ψ : ψ is subharmonic of U and lim sup U z→P ψ(z) ≤ f (P), ∀P ∈ ∂U} The set S is not empty because ∂U is compact and bounded below by some real constant m and therefore ψ(z) ≡ m is in S Define , for each z ∈ U u(z) = sup ψ∈S ψ(z). I claim that u solves the Dirichlet problem for f and U. We can show this in 3 steps. 1 The function u is bounded above. 2 The function u is harmonic. 62 / 73
  • 63. Perr´on Method Proof. Without loss of generality assume that f is real valued. First set S = {ψ : ψ is subharmonic of U and lim sup U z→P ψ(z) ≤ f (P), ∀P ∈ ∂U} The set S is not empty because ∂U is compact and bounded below by some real constant m and therefore ψ(z) ≡ m is in S Define , for each z ∈ U u(z) = sup ψ∈S ψ(z). I claim that u solves the Dirichlet problem for f and U. We can show this in 3 steps. 1 The function u is bounded above. 2 The function u is harmonic. 3 For each w ∈ ∂U, limU z→w u(z) = f (w). 63 / 73
  • 64. Perr´on Method Cont. 1 The function u is bounded above. Let M = maxζ∈∂U f (ζ), take ψ ∈ S and ε > 0, and let Eε = {z ∈ U : ψ(z) ≥ M + ε}. Assume that Eε is nonempty. 64 / 73
  • 65. Perr´on Method Cont. 1 The function u is bounded above. Let M = maxζ∈∂U f (ζ), take ψ ∈ S and ε > 0, and let Eε = {z ∈ U : ψ(z) ≥ M + ε}. Assume that Eε is nonempty. It can be shown that Eε is closed. 65 / 73
  • 66. Perr´on Method Cont. 1 The function u is bounded above. Let M = maxζ∈∂U f (ζ), take ψ ∈ S and ε > 0, and let Eε = {z ∈ U : ψ(z) ≥ M + ε}. Assume that Eε is nonempty. It can be shown that Eε is closed. Since Eε ⊂ U and U is bounded then Eε is bounded then Eε is compact. Since ψ is continuous and Eε is compact then then ψ attains its maximum at some point P ∈ Eε. Since the max is at least M + ε on Eε and ψ is less than M + ε outside Eε then ψ(P) is a max for ψ on all of U. Therefore by the maximum principle ψ is constant. Since P ∈ Eε then the constant must be greater than or equal to M + ε which contradicts ψ ∈ S. So Eε must be empty and thus ψ(ζ) ≤ M for all ζ ∈ U. Therefore u is bounded above. 66 / 73
  • 67. Perr´on Method Cont. 2 The function u is harmonic Let D(P, r) ⊆ U and p ∈ D(P, r). By definition of u there exists a sequence {ψj }∞ j=1 ⊆ S such that limj→∞ ψj (p) = u(p). Let Ψn(z) = max{ψ1(z), . . . , ψn(z)} for each z ∈ U. By properties of subharmonic functions Ψn(z) is subharmonic on U and Ψ1 ≤ Ψ2 ≤ . . . . Let Φn(z) = Ψn(z) if z ∈ U D(P, r) Poisson integral of Ψn ∂D(P,r) if z ∈ D(P, r) By the submean-value property the function Φn is subharmonic on U and also Φn ∈ S. Furthermore Φ1 ≤ Φ2 ≤ . . . on U by maximum principle. 67 / 73
  • 68. Perr´on Method Cont. Finally, ψn(p) ≤ Ψn(p) ≤ Φn(p) ≤ u(p) (∗) for each n and thus limn→∞ Φn(p) = u(p) By Harnack’s principle the function {Φn} converse to a harmonic function Φ on the disk D(P, r). Moreover Φ(P) = u(P) by (∗). Take a point q ∈ D(P, r), there is a sequence {ρj }∞ j=1 ⊆ S such that limj→∞ ρj (q) = u(q). Let ˜ρj (z) = max{ρj (z), ψj (z)} for z ∈ U and let Λn(z) = max{˜ρ1(z), . . . , ˜ρn(z)} for each z ∈ U. Again Λn is a subharmonic function on U with Λ1 ≤ Λ2 ≤ . . . and Λn(q) → u(q). 68 / 73
  • 69. Perr´on Method Cont. Define Hn(z) in the same way we defined Φn(z) before Hn(z) = Λn(z) if z ∈ U D(P, r) Poisson integral of Λn ∂D(P,r) if z ∈ D(P, r) {Hn} has a harmonic limit H on D(P, r) and H(q) = u(q). By design, Φ(z) ≤ H(z) ≤ u(z) on D(P, r) Hence H(p) = u(p) because Φ(p) = u(p). The harmonic function Φ − H is such that Φ − H ≤ 0 on ∂D(P, r) and (Φ − H)(p) = 0. The maximum principle says then that Φ ≡ H. Thus for any q ∈ D(P, r) the function H agrees with the function Φ on p. Thus u(p) agrees with u(q) for each point in D(P, r). Since P and r are chosen at random then we can conclude that u is harmonic. 69 / 73
  • 70. Perr´on Method Cont. 3 For each w ∈ ∂U, limU z→w u(z) = f (w). Fix w ∈ ∂U. For ε, k > 0 we have that g(z) = f (w) − ε + kbw (z) is subharmonic for z ∈ U. Choose δ > 0, a function of ε, so that f (z) > f (w) − ε whenever z ∈ ∂U ∩ D(w, δ). Thus g(z) ≤ f (z) for z ∈ ∂U ∩ D(w, δ) Note that bp(z) has a strictly negative upper bound on ∂U D(w, δ). Choose k, a function of ε, large enough such that g(z) ≤ f (z) for z ∈ ∂U D(w, δ) Thus g ∈ S, by the maximum principle for subharmonic functions, so g ≤ u and f (w) − ε = lim U z→w g(z) ≤ lim inf U z→w u(z) 70 / 73
  • 71. Perr´on Method Cont. By a similar argument lim supU z→w u(z) ≤ f (w) + ε. Thus f (w) − ε ≤ lim inf U z→w u(z) ≤ lim sup U z→w u(z) ≤ f (w) + ε Therefore we have the desired result limU z→w u(z) = f (w) : for w ∈ ∂U. The final thing to show is that u is unique. This follows from the maximum principle which says that the maximum for a harmonic function occurs on the boundary. 71 / 73
  • 72. References Greene, Robert E., and Krantz, Steven G. Function Theory of One Complex Variable. 3rd ed. 40. American Mathematical Society, 2006. Print. 72 / 73
  • 73. Thank You Thank you all for coming. I would like to give a special thank you to Dr. Edoh Amiran 73 / 73