Professional Documents
Culture Documents
Abstract
In this project, we study some results and theorems on the existence and uniqueness of
solutions for boundary value problems of the wave equation, the heat equation and the
Laplaces equation.
Contents
Introduction
1 The
1.1
1.2
1.3
1.4
1.5
1.6
1.7
2 The
2.1
2.2
2.3
2.4
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2
2
3
4
5
6
6
6
7
9
9
9
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
11
11
11
12
13
13
14
14
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
15
15
16
16
16
16
16
2.5
Heat Equation
The Maximum Principle . . . . . . . . .
The Fundamental Solution . . . . . . . .
The Heat Equation on R . . . . . . . . .
The Heat on half-line . . . . . . . . . .
2.4.1 The Dirichlet Problem on (0, )
2.4.2 The Neumann Problem on (0, )
Uniqueness . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
Conclusion
.
.
.
.
.
.
.
.
.
.
.
.
.
.
18
ii
Introduction
It is well known that most of the physical and engineering phenomena can be described in
mathematical models consisting of second order linear partial differential equations (PDEs).
The second order linear PDEs can be classified into three types, hyperbolic equations,
parabolic equations and elliptic equations. All the three types of equations can be reduced
to canonical forms. Hyperbolic equations reduce to a form coinciding with the wave equation
in the leading terms, the parabolic equations reduce to a form modeled by the heat equation,
and the Laplaces equation models the canonical form of elliptic equations.
Thus, the wave, heat and Laplaces equations serve as canonical models for all second
order constant coefficient PDEs. For this we need to study some results and theorems on
the existence and uniqueness of the solutions for the boundary value problems of the wave
equation, the heat equation and the Laplaces equation.
The first chapter treats the one-dimensional wave equation and models sound waves,
water waves, vibrations in solids, and torsional vibrations in a rod. The second chapter treats
the one-dimensional heat equation (Diffusion). The last chapter treats the two-dimensional
Laplace equation and the Poissons equation.
Chapter 1
The One-Dimensional Wave Equation
In this chapter we study the wave equation on the whole real line
2
2u
2 u
c
= 0, < x < ,
t2
x2
(1.1)
where c is the velocity of wave propagation along the string. This is the one-dimensional
wave equation which models sound waves, water waves, vibrations in solids, and torsional
vibrations in a rod. More details about the derivation of the wave equation can be found in
[1, 2].
1.1
c
t
x
+c
u = 0, < x < .
t
x
(1.2)
Theorem 1.1.1.
The general solution to the wave equation is given by
u(x, t) = f (x + ct) + g(x ct),
(1.3)
u
u u
u u
=
+
=
+
.
x
x x
c
t
x
+c
u = 2c
t
x
2c
u = 0,
1.2
The initial-value problem is to solve the wave equation with initial conditions
c2 uxx (x, t) = 0,
< x < , t > 0,
u(x, 0) = (x), ut (x, 0) = (x),
tt (x, t)
(1.4)
where and are two arbitrary functions of x. u(x, 0) = (x) defines the initial position of
the string, while ut (x, 0) = (x) is the initial velocity.
Theorem 1.2.1 (DAlembert Solution 1746).
The unique solution of the initial-value problem (1.4) is given by
1
1 Z x+ct
u(x, t) = ((x + ct) + (x ct)) +
( )d
2
2c xct
(1.5)
Proof.
First, setting t = 0 in Eq.(1.3), we get
(x) = f (x) + g(x),
(1.6)
(1.7)
f0 =
0 +
,
2
c
!
1
0
0
g =
.
2
c
Integrating, we get
f (s)
1 s
= 12 (s) + 2c
0 ( )d + A,
R
g(s) = 1 (s) 1 s ( )d + B,
2
2c 0
1.3
(1.8)
with f (x, t) is a given function. For instance, f (x, t) could be interpreted as an external
force acting on an infinitely long vibrating string.
Theorem 1.3.1.
The unique solution of problem(1.8) is given as:
1 Z x+ct
1 Z t Z x+c(t )
1
f (y, )dyd . (1.9)
( )d +
u(x, t) = ((x + ct) + (x ct)) +
2
2c xct
2c 0 xc(t )
Proof.
Let 4 be a triangle bounded domain with smooth boundary in the xt-plan with vertices
at (x0 , t0 ), (x0 ct0 , 0) and (x0 + ct0 , 0), and let 4 be a boundary of 4 such that 4 =
1 + 2 + 3 , (see figure 1).
Thus,
ZZ
f dtdx =
ZZ
4
ZZ
4
Z
4
Z
1
On 1 :
(c2 ux dt ut dx)
(c2 ux dt ut dx) +
t=0
Z
1
On 2 :
Z
2
Z
2
(c2 ux dt ut dx) +
Z
3
(c2 ux dt ut dx).
dt = 0. Thus,
(c ux dt ut dx) =
x + ct = x0 + ct0
(c2 ux dt ut dx) = c
=c
Z
2
Z
1
ut (x, 0)dx =
Z x0 +ct0
(x)dx.
x0 ct0
dx + cdt = 0. Thus,
(ux dx + ut dt)
(the total derivative of u(x, t)) ,
du
On 3 :
x ct = x0 ct0
Z
3
dx cdt = 0. Thus,
(c ux dt ut dx) = c
Z
3
= c
(ux dx ut dt)
du
Z x0 +ct0
(x)dx.
x0 ct0
Consequently, we have
1
1 Z x+ct
1 Z t Z x+c(t )
f (y, )dyd .
u(x, t) = ((x + ct) + (x ct)) +
( )d +
2
2c xct
2c 0 xc(t )
1.4
u(x, t) = f (x),
u(x, t) = g(x),
(1.10)
Theorem 1.4.1.
The unique solution of Goursats problem (1.10) is given by
x + ct
s+v sv
x ct
1 Z x+ct Z xct
u(x, t) = f
h
,
+g
u(0, 0) 2
dvds.
2
2
4c 0
2
2
0
Proof.
Here, the conditions are given on characteristics,(see figure 2).
2 2u
4c
= h( +
,
),
< + < , ,
2
2c
u(, 0) = f ( 2 ), u(0, ) = g( 2 ),
(1.11)
1 Z Z
+
+g
u(0, 0) 2
h
,
dvds.
2
2
4c 0 0
2
2
u(, ) = f
u(x, t) = f
x + ct
s+v sv
x ct
1 Z x+ct Z xct
h
,
+g
u(0, 0) 2
dvds.
2
2
4c 0
2
2
0
1.5
Now we try the same kind of problem for the wave equation.
1.5.1
= 0,
0 < x < , t > 0,
u(x, 0) = (x), ut (x, 0) = (x),
u(0, t) = 0.
(1.12)
Proceeding as before,
1
f (x) = (x) +
2
1
g(x) = (x)
2
1 Zx
( )d + A,
2c Z0
1 x
( )d + B, with A + B = 0.
2c 0
We must know the function g(x ct) for x ct < 0. For this taking into account that
the boundary condition
u(0, t) = 0, t 0.
f (ct) + g(ct) = 0 or g(y) = f (y), y < 0,
that is,
g(x ct) = f ((x ct)) = f (ct x), which defined on R .
Then, the solution is given by
u(x, t) =
1.5.2
x > ct,
x < ct.
= 0,
0 < x < , t > 0,
u(x, 0) = (x), ut (x, 0) = (x),
ux (0, t) = 0.
(1.13)
Again, for x ct we use the boundary condition ux (0, t) = 0, and from Eq.(1.3), we
obtain
f 0 (ct) + g 0 (ct) = 0.
Now, setting y = ct, we obtain
g 0 (y) = f 0 (y), y < 0.
So that,
g(y) = f (y) + C.
Therefore,
1 Z ctx
1
( )d + A + C.
g(x ct) = f (ct x) + C = (ct x) +
2
2c 0
Since,
1
B + (0) = g(0) = f (0) + C =
2
we have A + C = B and
1
(0) + A + C,
2
A + B = 0.
R
1 ((x + ct) + (x ct)) + 1 x+ct ( )d,
2
2c xct
R x+ct
1 ((x + ct) + (ct x)) + 1
( )d
2
1.6
2c
R ctx
0
x > ct,
( )d , x < ct.
Weak Solutions
Let H be the space of all functions defined on R2 and of class C2 which are zero outside
a bounded set of R2 .
Theorem 1.6.1.
Let u C2 , then
ZZ
"
"
ZZ
2
2
2u
2
2
2 u
c
dxdt
=
(x,
t)
c
dxdt
u(x, t)
t2
x2
t2
x2
R2
R2
Proof.
Let H,there exists a rectangle = (a, a) (b, b) such that is zero outside , (see
figure 3).
Let
ZZ
"
ZZ
ZZ
2
2
2
2
2
2
c
dxdt
=
dxdt
c
dxdt
u(x, t)
u(x,
t)
u(x,
t)
t2
x2
t2
x2
But,
Z a Z b
2
2
u(x, t) 2 dtdx.
u(x, t) 2 dxdt =
t
t
a b
Integrating by parts with respect to t from b to b two times and taking into account
(x, t) H, we obtain:
ZZ
"
u(x, t) 2 dt = u(x, t)
t
t
b
Z b
"
= u(x, t)
t
=
Z b
#b
Z b
b
#b
u
dt
t t
#b
"
(x, t)
t
b
+
b
Z b
(x, t)
2u
dt
t2
(x, t)
u
dt.
t2
Thus,
ZZ
Z a Z b
ZZ
2
2u
2u
u(x, t) 2 dtdx =
(x, t) 2 dtdx =
(x, t) 2 dtdx.
t
t
t
a b
Similarly,
Z b Z a
2
2
u(x, t) 2 dxdt =
u(x, t) 2 dxdt.
x
x
b a
Integrating by parts with respect to x from a to a,
ZZ
ZZ
Thus
ZZ
Z b Z a
ZZ
2u
2
2u
(x, t) 2 dxdt =
u(x, t) 2 dxdt =
(x, t) 2 dxdt.
x
x
x
b a
ZZ
2
2
2
2u
2
2 u
u(x, t)
c
dtdx
=
(x,
t)
c
dtdx.
t2
x2
t2
x2
Therefore
ZZ
ZZ
2
2
2
2u
2
2 u
u(x, t)
(x,
t)
c
dtdx
=
c
dtdx.
t2
x2
t2
x2
R2
R2
Remark: If u is a solution of utt c2 uxx = 0 then,
RR
u (tt c2 xx ) dtdx = 0.
Corollary 1.6.1.
For the wave equation
utt c2 uxx = 0.
A function of class C2 is a solution of Eq.(1.14) if and only if
ZZ
2
2
2
c
dtdx = 0, H.
u(x, t)
t2
x2
R2
Definition 1.6.1.
A solution of Eq.(1.14), such that
ZZ
2
2
2
u(x, t)
c
dtdx = 0, H,
t2
x2
R2
(1.14)
1.7
1.7.1
Let u be a solution of the wave equation (1.1). By multiplying both sides of Eq.(1.1) by
and integrating with respect to x, over I = [a, b] R, we obtain
Z b
a
Z b
u 2 u
u 2 u
2
dx
c
dx = 0.
t t2
a t x2
u
t
(1.15)
Since
!2
u 2 u
1 u
,
=
2
t t
2 t t
!
u 2 u
u 2 u
u u
,
=
t x2
x x t
x xt
u 2 u
1
=
x xt
2 t
u
x
!2
1 Z b u
2 a t
t
!2
+c
Let
u
x
1 Z b u
E(t) =
2 a
t
!2
2 u
x=b
u
dx = c
.
x t x=a
!2
+c
u
x
(1.16)
!2
(1.17)
dx.
1 Z b u
dE
=
dt
2 a t
t
1.7.2
!2
+c
u
x
!2
2 u
x=b
u
dx = c
.
x t x=a
Uniqueness
= f (x, t),
a < x < b, t > 0,
u(x, 0) = (x), ut (x, 0) = (x),
(1.18)
dE
= c2
v(b, t) v(b, t)
v(a, t) v(a, t) = 0,
dt
x
t
x
t
where E is the energy integral of v that is
!2
1Z b
E(t) =
v(x, t)
2 a
t
v(x, t)
x
+ c2
!2
dx
Thus
It follows
!2
1Z b
v(x, t)
E(t) =
2 a t
v(x, 0)
t
= 0, evaluating E(0)
!2
!2
1Z b
E(t) =
v(x, t)
2 a
t
= A,
v(x, t)
x
+ c2
dx
= 0.
!2
c2 Z b
dx +
v(x, t) dx = 0.
2 a x
Hence,
v(x, t) = 0
t
and
v(x, t) = 0,
x
that is
v(x, t) = B.
Using v(x, 0) = 0, we obtain B = 0. This means that
v(x, t) = 0 u1 (x, t) = u2 (x, t).
10
Chapter 2
The Heat Equation
In this chapter, we begin our study of the one dimensional heat equation (Diffusion)
u
2u
k 2 = 0,
t
x
a x b,
0 t T,
(2.1)
where k is called the thermal diffusivity. The general parabolic equation can be stated as
follows: Find a general solution u(x, t) of Eq.(2.1) which satisfies
(i) IC : the value of u at t = 0 is a known function.
(ii) BC : we are given for all t the value of u and
u
x
at x = a and x = b.
The solution (2.1) in the rectangle = (a, b) (0, T ) attains its maximum or minimum on
= (x = a, 0 t T ) (x = b, 0 t T ) (t = 0, a x b) .
2.1
on \ ,
if u1 u2 on , then u1 u2 on .
2. u(x, t) attains its extermum on , if m u(x, t) M on , then m u(x, t) M
on .
In particular, if u(x, t) = 0 on , then u(x, t) = 0 on .
2.2
Lemma 2.2.1.
If a > 0, then
Z
au2
du =
11
.
a
Proof.
2
I =
ax2
dx
ay 2
Z 2 Z
dy =
I=
rear drd =
,
a
.
a
2.3
(2.2)
Theorem 2.3.1.
The unique solution of problem (2.2) is given by
1 Z (xy)2
u(x, t) =
e 4kt (y)dy,
2 kt
t > 0.
(2.3)
Proof.
let ubt (, t) and ubxx (, t) be the Fourier transform of ut (x, t) and uxx (x, t) respectively. Thus
ubt (, t) =
ut (x, t)eix dx
Z
u(x, t)eix dx
=
t
= ub(, t),
t
we impose, u(x, t) 0 as x
ubxx (, t) =
=
u(x, t)eix
+ i u(x, t)eix
+ i
u(x, t)eix dx
= 2 ub(, t),
taking the Fourier transform to both sides of (2.2) we obtain an ODE in t
u
b(, t) + k 2 u
b(, t)
t
u
b(, 0)
= 0, t > 0,
b
= (),
with
2
k t
b
ub(, t) = ()e
.
12
2
b
Now let (,
t) = ek t we try to find (x, t) by Fourier inverse
1 Z b
(, t)eix d
(x, t) =
2
1 Z kt2 +ix
=
e
d
2
i
h
ix 2
ix 2
1 Z kt ( 2kt
) ( 2kt
)
=
e
d
2
ix 2
x2
1 Z kt( 2kt
) e 4kt
=
e
d
2
x2
1
e 4kt
=
( Lemma (2.2.1)).
2 kt
Now by convolution we get the result
u(x, t) =
(x y, t)(y)dy
1 Z (xy)2
=
e 4kt (y)dy.
2 kt
2.4
2.4.1
The problem is
u(x, 0) = (x),
u(0, t) = 0.
Theorem 2.4.1.
The unique solution of problem (2.4) is given by
(x+y)2
1 Z (xy)2
u(x, t) =
e 4kt e 4kt
(y)dy,
2 kt 0
t > 0.
Proof.
Extend to R
(x),
x > 0,
odd (x) = 0,
x = 0,
(x), x < 0.
Then the solution on R is
u(x, t) =
=
=
=
Z 0
Z 0
(x y, t)odd (y)dy
(x y, t)odd (y)dy +
Z
0
(x y, t)((y))dy +
(x y, t)odd (y)dy
(x y, t)(y)dy,
0
(x+y)2
1 Z (xy)2
e 4kt e 4kt
(y)dy.
=
2 kt 0
13
(2.5)
2.4.2
u(x, 0) = (x),
ux (0, t) = 0.
Theorem 2.4.2.
The unique solution of problem (2.6) is given by
(x+y)2
1 Z (xy)2
e 4kt + e 4kt
(y)dy,
u(x, t) =
2 kt 0
t > 0.
(2.7)
Proof.
Extend to R
(x),
x > 0,
even (x) = 0,
x = 0,
(x), x < 0.
Then the solution on R is
u(x, t) =
=
=
=
Z 0
Z 0
(x y, t)odd (y)dy
(x y, t)even (y)dy +
(x y, t)((y))dy +
Z
0
(x y, t)even (y)dy
(x y, t)(y)dy
0
(x+y)2
1 Z (xy)2
e 4kt + e 4kt
(y)dy.
=
2 kt 0
2.5
Uniqueness
The maximum principle can be use to proof for the uniqueness solution of the Dirichlet
problem.
Theorem 2.5.1.
If u1 and u2 are two solutions of the Dirichlet problem for the heat equation
= h(x, t),
u(x, 0) = (x),
u(a, t) = f (x), u(b, t) = g(x),
a x b, t > 0,
a x b,
(2.8)
then u1 = u2 .
Proof.
Let v(x, t) = u1 (x, t) u2 (x, t) be a solution of problem (2.8), then we have vt (x, t)
kvx x(x, t) = 0, v(x, 0) = 0, v(a, t) = 0, v(b, t) = 0 this means v(x, t) = 0 on , by the
maximum principle v(x, t) = 0 on ,
so that u1 (x, t) = u2 (x, t), a x b, t 0.
14
Chapter 3
The Laplaces Equation
Laplaces equation, also called the potential equation may be written as
2u 2u
+
= 0.
x2 y 2
(3.1)
2u
= 0,
2
i=1 xi
are the three and n-dimensional Laplace equation respectively. For Laplaces equation on a
region , a complete well-posed problem consist of the potential equation with appropriate
boundary conditions on .
3.1
=
x2
x
u r u
+
r x x
!2
2 u 2 r u 2 u
+
+
+
r2 x2 r x2
x
!2
r
x
2u
.
2
2u
=
y 2
r
y
!2
2 u 2 r u 2 u
+
+
+
r2 y 2 r y 2
y
!2
2u
.
2
Similarly,
We have
2 u 2 u r
+
=
x2 y 2
x
"
!2
r
+
y
!2
2 2 u
+
+
+
x2 y 2
x
= 0.
15
"
2u
2 r 2 r u
+
+
r2
x2 y 2 r
!2
+
y
!2
2u
2
x2 + y 2 and = tan1
y
x
, to get
r
x
2r
y
2r
y2
r
x2
q
q
= 2
,
=
,
=
,
=
,
x
x + y 2 x2
x2 + y 2 y 2
(x2 + y 2 )3 y
(x2 + y 2 )3
y
x
2
2xy
2
2xy
= 2
= 2
, 2 = 2
,
, 2 = 2
.
2
2
2
2
x
x + y x
(x + y ) y
x + y y
(x + y 2 )2
We obtain the Laplaces equation in a polar coordinates
2u
1
1 u
+ 2
+
2
r
r r
r
3.2
3.2.1
2u
= 0,
2
0 2, 0 < r < .
(3.2)
= 0,
on ,
u(a, y) = (y), u(b, y) = (x), c < y < d,
3.2.2
(3.3)
= 0,
on ,
ux (a, y) = (y), ux (b, y) = (x), c < y < d,
3.3
(3.4)
3.4
Uniqueness
Theorem 3.4.1.
Let = (a, b) (c, d) and if u1 and u2 be two solutions of the following problem
2
u
+ yu2 = 0, on \ ,
u(x, y) = 0,
on .
x2
(3.5)
Proof.
Let v(x, y) = u1 (x, y) u2 (x, y) be solution of problem (3.5), then by multiplying both sides
2v
2v
of x
2 + y 2 = 0 by v and integrating over ,
#
"
2v 2v
0=
+
dxdy
v
x2 y 2
#
#
Z d "Z b
Z b "Z d
2v
2v
v 2 dx dy +
v 2 dy dx
=
x
y
a
c
c
a
ZZ
=
that is
ZZ
ZZ
!2
!2
!2
v
+
y
v
+
y
dxdy
!2
dxdy
Z d
c
Z d
c
Z b
v
v
v dx,
v dy +
x a
y c
a
b
Z b
v
v
v dx.
v dy +
x a
y c
a
hence
ZZ
Z b
v
v
v dy +
v dx = 0,
x a
y c
a
!2
v
+
y
!2
dxdy
= 0,
we find
v
v
= 0 and
= 0, implies v(x, y) = c,
x
y
c = 0 because v(x, y) = 0 on .
Therefore,
v(x, y) = u1 (x, y) u2 (x, y) = 0
(3.6)
Theorem 3.4.2.
The problem (3.6) has one and only one solution on .
Proof.
Let v(x, y) = u1 (x, y) u1 (x, y) be solution of problem (3.6) then we have
2
v
v
+ y
on \ ,
2 = f f = 0,
v(x, y) = g g = 0,
on .
x2
17
Conclusion
We have considered the wave, heat and Laplaces equations. We have demonstrated that
the exact solution can be obtained in a straightforward manner by using a direct method.
Also, theorems on the uniqueness solutions were obtained.
18
Bibliography
[1] Walter, A.Stranss, Partial Differential Equations, An Introduction, Wiley, 1992
[2] T.Hillen,I.Eleonard and H.Van Roessel, Partial Differential Equations, Theory and Completely Solved Problem, Wiley, 2012
19