Вы находитесь на странице: 1из 37

124A PARTIAL DIFFERENTIAL EQUATIONS

DENIS A. LABUTIN
This is the outline for the course. We emphasize the key ideas and the structure,
and often omit the details. The details will be covered on the lectures. They also
can be found in the textbook (W. A. Strauss PDEs, an Introduction, 2nd ed.) and
in the extended UCSB 124-course notes by Viktor Grigoryan. The two sources
contain much more than will be covered during the quarter.
1. Basics
(1) A partial dierential equation (in two independent variables) is a relation
F(x, y, u, u
x
, u
y
, u
xx
, u
xy
, . . .) = 0
for the unknown function u(x, y) . The standard agreement is to denote
u
x
=
x
u = u
x
.
All laws in sciences are formulated in the form of (system of) PDEs. A
function u is a solution of the PDE in a region if
F(x, y, u(x, y), u
x
(x, y), u
y
(x, y), u
xx
(x, y), u
xy
(x, y), . . .) 0
for (x, y) . Thus it is very easy to check if a given function is a solution
of the given PDE.
(2) The general solution to the given PDE is the formula for all solutions. For
example, we know from the calculus that the general solution u(x) of
du
dx
u = 0
is
u = Ce
x
,
where C is any constant.
Another example: we will learn later that the general solution u(x, y)
to the equation
u
tt
u
xx
= 0
is given by
u(t, x) = f(x t) +g(x +t),
where f, g are arbitrary functions of one variable. This means that all
solutions of the equation are given by this formula. It is impossible to
derive this fact now, we need to develop some techniques for this.
Date: June 25, 2014.
1
2 DENIS A. LABUTIN
(3) Calculus technique for solving simple PDEs.
(4) Some simple PDEs can be integrated using just the calculus technique. In
order to successfully solve the problems of this type one needs to understand
the following argument: according to the basic calculus the general solution
z(x) to the simplest equation
dz
dx
= 0
is
z(x) = C,
where C is any constant. But this means that the general solution u(x, y)
to the equation
u
x
= 0
is
u(x, y) = C(y),
where C(y) is any constant with respect to x , which means that it is any
function of y . When writing the solutions down it is better not to use C
to denote functions, and keep it for constants only. Thus a better way of
writing is
u
x
= 0 u = (y),
where is any function of one variable.
(5) Similarly, the basic calculus tells us that the solution z(x) of z

= f(x)
is
z

= f(x) z =
_
f(x) dx +C,
where C is any constant. This means that
u
x
= f(x, y) u =
_
f(x, y) dx +(y),
where is any function of one variable. Below are some examples of
calculus technique solutions to PDEs.
(6) Solve (this means nd the general solution to) the equation
u
x
= e
2x+y
.
There is no /y in the left hand side, and no u in the right hand side.
Therefore the solution u(x, y) is given by
u =
_
e
2x+y
dx +(y)
= e
y
_
e
2x
dx +(y)
= e
y
1
2
e
2x
+(y).
Thus the answer is
u =
e
2x+y
2
+(y),
where is any function.
124A PARTIAL DIFFERENTIAL EQUATIONS 3
(7) Here is one more example of the calculus technique for solving simple PDEs.
Find all solutions u(x, y) , v(x, y) to the equations
u
x
+u = 0, v
xx
+v = 0.
Using the calculus separation of variables we see that the equation
z

+z = 0
has the general solution
z(x) = Ce
x
.
The equation for u is the same equation. There is no /y and the
variable y enters only as a parameter. Hence
u
x
+u = 0 u = (y)e
x
,
where is any function.
Using the calculus separation of variables we see that the equation
z

+z = 0
has the general solution
z(x) = C
1
cos x +C
2
sinx,
where C
1,2
are arbitrary constants. The equation for v is the same
equation. There is no /y and the variable y enters only as a parameter.
Hence
v
xx
+v = 0 v = (y) cos x +(y) sinx,
where , are arbitrary functions.
(8) Frequently there is a need to apply several calculus steps. Here is an ex-
ample of such problem. Solve
u
xy
= 0.
Denote
v = u
y
.
Then
u
xy
= 0 v
x
= 0 v = (y),
where is any function of one variable. Next,
u
y
= (y) u =
_
(y) dy +(x),
where , are any functions of one variable. Since is any, then
_
(y) dy
is also any function of one variable. Thus the answer is
u = (y) +(x),
where , are any functions of one variable.
Notice that , are the functions of one variable, and u(x, y) = (y)+
(x) is a function of two independent variables of a very special form. For
example (x)(x) is not of this form. Hence x + e
y
is a solution, but
xe
y
is not.
4 DENIS A. LABUTIN
(9) The integrating factor technique from calculus will also be used in the
course. Equation for z(t)
dz
dt
+a(t)z = f(t)
can be integrated using the following observation. Multiply the equation
by the integrating factor
e
R
a(t) dt
to discover that
e
R
a(t) dt
dz
dt
+e
R
a(t) dt
a(t)z = e
R
a(t) dt
f(t).
The point here is that now
e
R
a(t) dt
dz
dt
+e
R
a(t) dt
a(t)z =
d
dt
_
e
R
a(t) dt
z
_
.
Hence the equation becomes
d
dt
_

_
= g(t)
which can be integrated.
Problem 1. Find the general solution of
u
xy
+a(x, y)u
x
= 0
in a neighborhood of (0, 0) . Here a(x, y) is a given function.
Set
v = u
x
to discover at once that
v
y
+a(x, y)v = 0.
This is an ODE in the y -variable, and x enters as a parameter. Separat-
ing the variables discover
dv
v
= a(x, y) dy
and nd using the integrating factor that
ln[v(x, y)[ =e

R
y
0
a(x,) d+C(x)
v = (x)e

R
y
0
a(x,) d
with an arbitrary function . Returning to u discover that
u
x
= (x)e

R
y
0
a(x,) d
.
This is just an ODE in the x -variable, and y enters as a parameter.
Integrating we nd
u(x, y) =
_
x
0
()e

R
y
0
a(,) d
d +(y),
where , are arbitrary functions of one variable.
Problem 2. Find the general solution of
u
xy

1
x y
u
x
+
1
x y
u
y
= 0.
Hint: use the observation that
(x y)vx + v =

(x y)v

x
, (x y)vy v =

(x y)v

y
.
124A PARTIAL DIFFERENTIAL EQUATIONS 5
(10) Operators and equations. Let u be any function. A dierential op-
erator is an expression involving u and its derivatives. For example, the
Monge-Ampere operator is
det D
2
u = u
xx
u
yy
u
2
xy
,
the Laplace operator is
u = u
xx
+u
yy
,
the wave operator is
u = u
tt
u
xx
,
c
u = u
tt
c
2
u
xx
,
the heat operator is
u
t
u
xx
,
the transport operator is
u
x
+u
y
,
the Schroedinger operator is
1

1
u
t
u
xx
.
Operators a usually denoted as Au , or A[u] , or A(u) , or Lu , ... .
(11) The most important class of the operators is the class of linear operators.
Operator L is called linear if the following two conditions hold:
L(u +v) = L(u) +L(v), L(Cu) = CL(u)
for all functions u, v and any constant C . To check if the operator is
linear we just need to verify these conditions.
(12) A PDE is called linear homogeneous if it can be written in the form
Lu = 0
with some linear operator L .
A PDE is called linear inhomogeneous if it can be written in the form
Lu = f(x, y)
with some linear operator L . The important part here is that in the right
hand side we have some function of x, y but the unknown u is absent.
(13) Our main objects of study will be the transport operator
(
t
+c
x
)u = u
t
+cu
x
, u = u(t, x),
the wave operator
u
def
=
_

2
t
+
2
x
_
u = u
tt
u
xx
, u = u(t, x),
the heat operator
(
t

2
x
)u = u
t
u
xx
, u = u(t, x),
and the Laplace operator
u
def
= (
2
x
+
2
y
)u = u
xx
+u
yy
, u = u(x, y).
Show that they are linear operators.
Linear equations can have coecients depending on x, y . For example,
show that the equation
u
x
+yu
y
= 0
6 DENIS A. LABUTIN
is a linear (homogeneous) equation.
(14) Equation
u
x
+uu
y
= 0
is nonlinear. Indeed, every term in it contains u . Hence the equation is
linear if and only if the operator
Au = u
x
+uu
y
is linear. But A is not linear because the two conditions of the linearity
are not satised. For example if C is a constant, then
A(Cu) = (Cu)
x
+ (Cu)(Cu)
y
= Cu
x
+C
2
uu
y
.
At the same time
CAu = Cu
x
+Cuu
y
.
Hence
A(Cu) ,= CA(u)
for all constants C .
(15) Superposition principle for linear equations. The key feature of lin-
ear equations is the superposition principle. It is also called the linearity
principle. It consists of two parts.
The rst part states that if u
1
, . . . , u
N
are solutions to the linear
homogeneous equation
Lu = 0,
then for any constants C
1
, . . . , C
N
the function C
1
u
1
+ +C
N
u
N
is
also a solution:
L(C
1
u
1
+ +C
N
u
N
) = 0.
This follows at once from the denition of the linearity of L .
The second part states that if u
0
solves the linear inhomogeneous equa-
tion
Lu = f,
and if v is the general solution to the homogeneous equation
Lv = 0
then the general solution to the inhomogeneous equation is given by
u = u
0
+v.
(16) The second part is the most important and useful for problem solving.
Indeed, in plain language is states that
all solutions to Lu = f = some solution to Lu = f
+all solutions to Lu = 0.
Thus to nd all solution to Lu = f it is enough to nd all solutions only
to Lu = 0 and somehow guess just one particular solution to Lu = f .
(17) Let us illustrate the power of the superposition principle by solving a prob-
lem. Solve (that is, nd all solutions to)
u
xy
= x
2
y.
124A PARTIAL DIFFERENTIAL EQUATIONS 7
This is a linear inhomogeneous equation. Hence we can apply the super-
position principle. We have computed earlier the general solution to the
homogeneous equation:
v
xy
= 0 v = f(x) +g(y),
where f, g any functions. Hence we just need to guess one solution to the
inhomogeneous equation. This is easy for our equation. For example
u
0
=
x
3
3
y
2
2
=
x
3
y
2
6
will do the job. Therefore, by the superposition principle the answer is:
u(x, y) =
x
3
y
2
6
+f(x) +g(y),
where f, g any functions.
(18) The superposition principle will be one of the main tools in this course both
for the theory and the problem solving. Superposition principle does not
hold for nonlinear equations.
2. Changing variables in PDEs
(1) The changing variables technique is fundamental for PDEs. It is based on
various tools from calculus such as chain rule, implicit dierentiation, and
so on. Many problems in this course require the ability to change variables
correctly. First we explain the generalities, underline the dicult part in
the change of variables procedure, and then solve some problems.
(2) To abbreviate we write x = (x
1
, . . . , x
n
) , = (
1
, . . . ,
n
) , u
x
=
(
x1
u, . . . ,
xn
u) , and so on. In the dierential equation
F(x, u, u
x
, . . .) = 0
we would like to change the variables x . The change of variables
x = H()
at once gives
v() = u(H()).
and
(2.1) F(H(), v(), (u
x
)(H()), . . .) = 0.
We know F and H . Consequently we almost have the dierential equa-
tion for the new unknown v() .
It is just left to relate (u
x
)(H()) with v

. This is the somewhat


complicated part. To accomplish that we need the inverse change
= (x).
The inverse implies
(H()) = , H((x)) = x.
8 DENIS A. LABUTIN
Therefore together with v() = u(H()) we also have u(x) = v((x)) .
By the chain rule
u
x
(x) =

v()

=(x)

x
(x),
and we need to express this in variables:
(u
x
)(H()) =
_

v()

=(x)

x
(x)
_

x=H()
= v

() (
x
) (H()).
We see that the main computation when changing variables is nding
(
x
) (H())
and the higher derivatives needed for u
xx
, u
xxx
, . . . . That is, we need
to express the derivative
x
(x) as a function of by substituting x =
H() .
When this is done, equation (2.1) becomes
0 = F(H(), v(), v

() (
x
) (H()), . . .)

0 = G(, v, v

, . . .) .
(3) Thus to change the independent variables x to in the equation
F(x, u, u
x
, . . .) = 0
we need:
Step 1. Use x = H() to introduce v() = u(H()) , and write the inverse
change = (x) so that u(x) = v((x)) .
Step 2. Compute by the chain rule

x
u(x) =

v((x))
x
(x)
substituting x = H()
(
x
u)(H()) =

v() (
x
)(H()) ,
and higher derivatives, if necessary

2
x
u(x) =
x
_

v((x))
x
(x)
_
=
2

v((x)) (
x
(x))
2
+

v((x))
2
x
(x)
substituting x = H()
(
xx
u)(H()) =
2

v()
_
(
x
)(H())
_
2
+ ;
Step 3. Plug the results of steps 1 and 2 into the equation to write it in the
new variables
0 = F(H(), u(H()), u
x
(H()) . . .)
= G(, v, v

, . . .).
(4) The successful strategy for changing variables is not to memorize the general
formula above, but to repeat the three steps for individual problem. Let us
show this by examples.
124A PARTIAL DIFFERENTIAL EQUATIONS 9
(5) In the eiconal equation
(u
x
)
2
+ (u
y
)
2
= 1
pass to the polar coordinates
x = r cos , y = r sin.
(When dealing with polar coordinates one always assumes (x, y) ,= (0, 0)
and hence r ,= 0 .)
First, changing the variables from (x, y) we obtain the new unkown
v(r, ) = u(x, y)

x=r cos , y=r sin


.
Second, to obtain the equation for v we need the inverse change
r = r(x, y), = (x, y).
Then u(x, y) = v(r(x, y), (x, y)) and
u
x
= v
r
r
x
+v

x
u
y
= v
r
r
y
+v

y
.
We need to express the right hand sides in terms of (r, ) and plug the
result into the eiconal equation.
With polar coordinates the inverse change does not have a convenient
form. The trick is to use the implicit dierentiation of polar coordinates:
x r(x, y) cos (x, y), y r(x, y) sin(x, y),
therefore
1 = r
x
cos r sin
x
0 = r
y
cos r sin
y
1 = r
y
sin +r cos
y
0 = r
x
sin +r cos
x
and
r
x
=cos
r
y
=sin

x
=
sin
r

y
=
cos
r
Combining the results discover
1 = u
2
x
+u
2
y
= (v
r
r
x
+v

x
)
2
+ (v
r
r
y
+v

y
)
2
=
_
v
r
cos +v

sin
r
__
2
+
_
v
r
sin +v

_
cos
r
__
2
= v
2
r
+
1
r
2
v
2

.
10 DENIS A. LABUTIN
3. Derivation of equations and boundary conditions
(1) The Gauss divergence theorem is crucial for deriving PDEs. It says that if
is the boundary of the domain , and if n is the outer normal on
, then
___

div

F dvol =
__

F ndA.
Here dvol is the volume lement in and dA is the area element on
.
(2) Derivation of the transport equation. Let be the density of the
transported component,

V be the velocity of the transporting ow.
Let also f be the density of the external sources. That is the external
production of the component at the point x = (x
1
, x
2
, x
3
) during the time
interval (t, t + t) is
M = f(t, x) dvol(x) t.
The key observation: let m be the mass transported by the ow

V
through the at plate S in the direction of the unit normal n , [n[ = 1 ,
during the time interval t . Then
[m[ = vol(R)
= S height
= S [

V [t cos .
Im
Re
&%
'$
&%
'$

q
1 1
q
Rez = 1/2
D = [z[ > 1 Rez = 1/2, Imz > 0
A
1
= 2e
i
r
r
A
2
= 2e
i
We must think that m > 0 if

V points in the same direction as n ,
and m < 0 if the directions are opposite. Hence
m = S t (

V n).
The mass balance condition must always hold for the component. There-
fore for any spacial volume with the boundary and the external
124A PARTIAL DIFFERENTIAL EQUATIONS 11
normal n we have
_

(t + t, x) dvol(x)
_

(t, x) dvol(x) =production inside


ux through in direction n
=t
_

f dvol
t
_

(t, x)

V ndA(x)
=t
_

f div
_
(t, x)

V
_
dvol(x).
Divide by t 0 to discover
_

t
(t, x) dvol(x) =
_

f div
_
(t, x)

V
_
dvol(x).
Then divide by vol() 0 to nd that the equation for the density
= (t, x)
of the component transported by the ow

V with the sources density f
is

t
= div(

V ) +f.
In real processes the presence of the component inuences the ow

V =

V (t, x, ) and the sources density f = f(t, x, ) . So the transport equation


is nonlinear:

t
+ div
x
((t, x)

V (t, x, (t, x))) = f(t, x, ).


If the ow and the source are not inuenced by the component the equation
becomes
(3.2)
t
+ div
x
((t, x)

V (t, x)) = f(t, x).


Prove that (3.2) is a linear (non-homogeneous) equation for the unknown
= (t, x) . If the ow has a constant velocity

V (t, x)

C,
the transport equation is

t
+

C = f.
If the constant velocity ow is one dimensional then the transport equation
becomes
= (t, x),
t
+c
x
= f(t, x).
(3) Derivation of the heat equation. The heat equation (also called the
heat transport equation) shares some features with the transport equation.
The unknown is the temperature T = T(t, x) . The key fact is the Fouriers
law of the heat conduction: let Q be the amount of the heat energy
transported through the at plate S in the direction of the unit normal
n , [n[ = 1 , during the time interval t . Then
Q = kt S gradT n.
Here k is the (positive) materials conductivity coecient. Notice the
similarities with the transport. What is the meaning of the " " sign in
the formula?
12 DENIS A. LABUTIN
Let f be the intensity of the external heat sources. That is the amount
of energy produced externally at the point x during the time interval
(t, t + t) is
E = f(t, x) dvol(x) t.
The heat energy E of the volume dvol is given by
E = C T dvol,
where C is the thermal capacity coecient of the material. Using the
Fourier law we write the conservation of energy for any spacial volume
_

CT(t + t, x) dvol(x)
_

CT(t, x) dvol(x) =heat production inside


heat ux through in direction n
=t
_

f dvol
+ t
_

k gradT ndA
=t
_

f + div
_
k gradT
_
dvol.
Divide by t 0 to discover
_

C
t
T(t, x) dvol(x) =
_

f + div
_
k gradT(t, x)
_
dvol(x).
Then divide by vol() 0 to nd that the equation for the temperature
T = T(t, x, y, z)
with the sources density f is

t
T =
k
C
div(gradT) +f.
(4) Laplace operator. The operator
div(gradu) = (
2
x
+
2
y
+
2
z
)u
is called the Laplace operator or Laplacian and is denoted by ,
= div(grad ) =
2
x
+
2
y
+
2
z
.
Thus the heat equation for the unknown T = T(t, x, y, z) with the given
sources of the density f is

t
T = aT +f.
In the one dimensional case the heat equation becomes
T = T(t, x),
t
T = a
2
x
T +f, a > 0.
Suppose that the temperature regime is stationary, that is T = T(x, y, z)
and

t
T = 0.
Then the temperature distribution is described by the Laplace (or Poisson)
equation
aT = f.
Solutions u of
u = 0
are called the harmonic functions.
124A PARTIAL DIFFERENTIAL EQUATIONS 13
(5) Diusion equation. This is the same equation as the heat equation. Let
N = N(t, x) be the concentration of the component which is being diused
in medium. The key for the derivation is the Ficks diusion law: let
be the amount of the component passed through the at plate S in the
direction of the unit normal n , [n[ = 1 , during the time interval t
under the natural diusion. Then
= dt S gradN n.
Here k is the (positive) materials diusion coecient. This is exactly the
same formula as in the Fouriers heat conduction law.
(6) Wave equation. We derive the equation for small vibrations of a one
dimensional horisontal string. "Small" means that if the length of the
string at rest is l , and if the change in length during the vibration is l ,
then
l
l
1.
This implies that the change in the tension force value
T
T
0

l
l
1.
Therefore the tension force has the constant value T
0
during the vibra-
tions. "Small" also means tag the pieces of the string move only in the
vertical direction. Accepting these observations we derive the equation.
Let u(t, x) be the vertical coordinate of a piece (x, x + x) of the
string. The Newtons law of the motion projected on the vertical axis
states
mu
tt
= F +T
vert
(t, x) +T
vert
(t, x + x).
Here F is an external vertical force on the piece (x, x+x) , and m =

0
x . For the tension force we have

T(x) = T
0
(tangent vector at x)
and

T(x + x) = T
0
tangent vector at x + x.
Here the tangent vector to the graph of u is
(1, u
x
)
_
1 +u
2
x
.
The smallness of vibrations allows to write
u
x
_
1 +u
2
x
= u
x
(1 +
1
2
u
2
x
+ ) u
x
.
Thus
T
vert
(t, x) +T
vert
(t, x + x) =T
0
(u
x
(t, x + x) u
x
(t, x)).
Therefore the Newtons law becomes
u
tt
=
T
0

0
u
x
(t, x + x) u
x
(t, x)
x
+
F

0
x
.
Letting x 0 discover
u
tt
= c
2
u
xx
+f(t, x),
where f is the given external force density and c =
_
T
0
/
0
. The oper-
ator

c
u =
_

2
t
c
2

2
x
_
u
is called the wave (or Dalamberts) operator.
14 DENIS A. LABUTIN
(7) Initial vaue problem. For the equation of the type

t
u +L(
x
)u = f
the initial value problem is to nd u(t, x) satisfying
_
u
t
+Lu = f for t > 0, x (, +)
u[
t=0
= (x).
Here f(t, x) and (x) are given functions.
For the wave equation the initial value problem is to nd u(t, x) satis-
fying
_

_
u
tt
+u
xx
= f for t > 0, x (, +)
u[
t=0
= (x)
u
t
[
t=0
= (x).
Here f(t, x) , (x) , and (x) are given functions.
(8) Boundary conditions. In one dimension u = u(t, x) , t > 0 , a < x <
b , the most common boundary conditions are:
Dirichlet, functions f(t) , g(t) are given, u satsies
u[
x=a
= f(t), u[
x=b
= g(t) for all t > 0;
Neumann, functions f(t) , g(t) are given, u satsies
u
x
[
x=a
= f(t), u
x
[
x=b
= g(t) for all t > 0;
Robin, positive constants C
1
, C
2
> 0 and functions f(t) , g(t) are given,
u satsies
(u
x
C
1
u)[
x=a
= f(t), (u
x
+C
2
u)[
x=b
= g(t) for all t > 0;
Conditions at , u and all its derivatives tend to 0 as x fast enough for all
t > 0 .
(9) The meaning of the Dirichlet conditions is obvious (prescribing the temper-
ature of the ends of the rod or the motion of the ends of the string).
The least obvious conditions are the Robins. For the heat propagation
they are derived by the following argument based on Newtons law of the
heat energy conduction and on Fouriers heat transport law. Suppose that
the right end of the rod, 0 < x < l , exchanges the energy with the massive
reservoir of the given temperature T(t) . The Newtons law states that
the energy ux Q from the rod into the reservoir in time t is
proportional to the temperature dierence,
Q = C(u(t, l) T(t))t.
At the same time
the amount of heat transported through the right end of the rod
also outside the rod in time t by Fouriers law is
Q = u
x
(t, l)t.
Hence we must have u
x
(t, l) = C(u(t, l) T(t)) , which is equivalent to
_
u
x
+
C

u
_

x=l
=
C

T(t).
Similarly, using the Fourier law at the left end of the rod we derive
Q = u
x
(t, 0)
124A PARTIAL DIFFERENTIAL EQUATIONS 15
and
_
u
x

u
_

x=0
=
C

T(t).
For the string vibration consider the situation when, say, the left end of
the string is attached to an elastic spring acting with the returning vertical
force
F = ku(t, 0).
Then the Newtons law for the element x projected on the vertical axis
states that
(x)u
tt
(t, 0) = T +F
= T
0
u
x
(t, 0) ku(t, 0).
(Here to get the formula T = T
0
u
x
we assumed that the vibrations are
small, hence t u
x
is small, and hence
sin tan = u
x
,
where is the angle at which the end of the string slides along the vertical
axis.) Letting x 0 deduce that
_
u
x

k
T
0
u
_

x=0
= 0.
The tension force on the right end of the string is
T = T
0
u
x
(t, l),
and the similar argument gives
_
u
x
+
k
T
0
u
_

x=l
= 0.
4. Transport and the first order equations
(1) Main tool. The important calculus notion for the transport equation is
the directional derivative of a function. Let u be a multivariable function,
u = u(r),
r = (r
1
, . . . , r
n
) . Let

V be a xed vector. Consider the line through the
point r
0
in the direction

V :
r(t) = r
0
+t

V , < t < .
Restrict our function u to this line. This restriction is just a function of
one variable
u(r(t)), < t < .
Now, the derivative of the function u at the point r
0
along the vector

V
is just the usual one dimensional derivative of u(r(t)) :
u(r
0
)

V
=
du(r
0
+t

V )
dt

t=0
.
16 DENIS A. LABUTIN
This is the denition of the derivative along the vector. Using the chain
rule we continue the computation, and obtain the important useful formula
for
u(r0)

V
:
du(r
0
+t

V )
dt

t=0
=
n

j=1
u(r
0
)
r
j
V
j
= u(r
0
)

V .
Thus
u(r
0
)

V
= u(r
0
)

V .
If
[

V [ = 1
then the derivative along

V is called the derivative in the direction

V , or
the directional derivative.
(2) Solution to the simplest transport equation. We nd the general
solution u(x, y) of the equation
au
x
+bu
y
= 0,
with the constant

V = (a, b) ,= 0 . This is an easy matter since
au
x
+bu
y
= u

V =
u

V
.
Thus the transport equation is equivalent to
u

V
= 0
_
u is constant along any line
parallel to

V
_
How to express the last property of u(x, y) as a formula?
One way to argue is as follows. Any line parallel to (a, b) has the
equation
bx +ay = C
with some constant C . Along this line the value of u is (some other)
constant. Therefore the level sets of functions u(x, y) and bx + ay
coincide! This is possible if and only if
u(x, y) = f(bx +ay)
with some function f of one variable. Thus
au
x
+bu
y
= 0 u(x, y) = f(bx +ay)
where f is an arbitrary function of one variable.
Here is another way to derive this formula. Namely, x a point (x
0
, y
0
) . The line
through (x
0
, y
0
) parallel to (a, b) has the equation
bx + ay = bx
0
+ ay
0
.
At least one coecients has to be nonzero, suppose a = 0 . Since u is constant along
this line
u(x
0
, y
0
) = u

x,
bx bx
0
+ ay
0
a

x=0
= u

0,
bx
0
+ ay
0
a

.
But u(0, z) is just a function of one variable z . Denote it by f(z) . Then
u(x
0
, y
0
) = f

bx
0
+ ay
0
a

= g(bx
0
+ ay
0
)
with the one variable function g , g(z) = f(z/a) .
124A PARTIAL DIFFERENTIAL EQUATIONS 17
(3) As we just saw, the lines bx +ay = c which are parallel to (a, b) , play
an important role for the transport equation au
x
+ bu
y
= 0 . They are
called the characteristics of the equation. Above it is explained, that the
equation is equivalent to the vanishing of the derivative of u along the
characteristic.
Using the formula for the general solution we discover at once that the
solution to the initial value problem
_
u
t
+cu
x
= 0 for t > 0, x (, +)
u[
t=0
= (x)
is given by the wave propagating to the right
u = (x ct).
Problem 3. Why this formula indeed gives a wave propagating to the
right? Draw pictures.
(4) Using the property of the characteristics it is easy to solve the initial value
problem for the inhomogeneous transport equation in the general form
_
u
t
+cu
x
= f(t, x) for t > 0, x (, +)
u[
t=0
= (x) for x (, +) .
However, it is not very convenient for problem solving. For that it is almost
always easier to compute the solution using the superposition principle.
Since the equation is linear by the superposition principle we write
u = v +w
where
_
v
t
+cv
x
= f for t > 0, x (, +)
v[
t=0
= 0 for x (, +) .
and
_
w
t
+cw
x
= 0 for t > 0, x (, +)
w[
t=0
= (x) for x (, +) .
We already know that
w(t, x) = (x ct).
It is left to nd v .
The key observation for that is that any parametrization of a straight line r =
r
0
+

V , according to the chain rule gives


v(r)

V
= v(r)

V =
d
d
v(r
0
+

V ).
That is, our PDE is equivalent to the ODE along the characteristic with

V = (1, c).
Now, x a point (t
0
, x
0
) and consider the characteristic line
x ct = x
0
ct
0
x = ct + (x
0
ct
0
)
passing through it. Take t as the parameter along the line. Then according to our
observation, the restriction z of u to the characteristic,
z(t) = v(t, x
0
ct
0
+ ct),
satises
8
<
:
dz
dt
= f(t, x
0
ct
0
+ ct) for t > 0
z(0) = 0,
18 DENIS A. LABUTIN
and the integration gives
v(t
0
, x
0
) = z(t
0
)
= z(0) +
Z
t
0
0
dz(t)
dt
dt
= 0 +
Z
t
0
0
f(t, x
0
ct
0
+ ct) dt.
Hence the answer is
u(t, x) = (x ct) +
Z
t
0
f(, x c(t ) d.
(5) Variable coecients rst order equation. The equation
a(x, y)u
x
+b(x, y)u
y
= 0
similarly to the constant coecients transport equation, can be rewritten
with the vector eld

A(x, y) =
_
a(x, y)
b(x, y)
_
as
a(x, y)u
x
+b(x, y)u
y
= 0
u


A(x, y)
= 0.
Unlike the constant coecients case, the vector eld now depends on
(x, y) . Instead of the characteristic lines we need the characteristic curves.
By denition those are the curves tangent to

A . Formally this means they
solve the system
_
x(t)
y(t)
_
=

A(x(t), y(t)).
Along any such curve
d
dt
u(x(t), y(t)) = au
x
+bu
y
and therefore
u


A
= 0
_
u is constant along any curve
tangent to

A
_
.
How to express the last condition as a formula?
The easiest way to argue is as follows. Suppose we have the equation of
the characteristic curves not in the form
t
_
x(t)
y(t)
_
,
but in the form of a level curve
F(x, y) = const.
The gradient F is orthogonal to the level curve, whereas

A is tangent
to the level curve. Hence
F(x, y)

A(x, y)
(that is aF
x
+ bF
y
= 0 ). At the same time for the unknown u we also
have
au
x
+bu
y
= 0 u

A = 0,
and hence
u(x, y)

A(x, y).
But u is orthogonal to the level curve of u . Therefore the tangent line
to the level curve of u coincides with the tangent line to the level curve
124A PARTIAL DIFFERENTIAL EQUATIONS 19
of F . But this means that the level curves of F and u coincide! This
is equivalent to
u(x, y) = f(F(x, y))
with an arbitrary function f of one variable.
(6) Thus for the problem solving we need to have an eective way of nding
the characteristic curve in the form
F(x, y) = const.
This is easy: for any curve (x(t), y(t)) expressed as a graph y = y(x) we
have
dy
dx
=
dy
dt
dx
dt
=
y
x
.
But the equation for the characteristics is
x = a(x, y)
y = b(x, y)
.
Hence
dy
dx
=
b(x, y)
a(x, y)
.
It is better to write this equation in more symmetric form
dy
b(x, y)
=
dx
a(x, y)
integration
F(x, y) = const.
This is the formula for the characteristic curves for the equation
au
x
+bu
y
= 0.
Then the general solution to the equation is
(4.3) u(x, y) = f(F(x, y)),
where f is an arbitrary function of one variable.
(7) Let us give a rigorous proofs of (4.3).
The equation
(4.4) a(x, y)ux + b(x, y)uy = 0
has the characteristic system
x = a(x, y)
y = b(x, y).
The key notion for writing down the formula for u is the rst integral of the charac-
teristic system.
Suppose we know that F is the rst integral of the system. This means that it is
constant along any solution (x(t), y(t)) . Equivalently
0 = tF(x(t), y(t))
= Fx x + Fy y
= Fxa + Fyb .
That is our F satises
aFx + bFy = 0.
Suppose Fy = 0 (if Fx = 0 the argument is similar). Introduce the the map m ,

x
y

x
F(x, y)

.
That is, make the change of variables (, ) = m(x, y) ,
= x
= F(x, y)
.
20 DENIS A. LABUTIN
Geometrically m maps vertical lines into vertical
{x = const}
m
{ = const},
and the level curves of F into horisontal
{(x, y) | F(x, y) = const}
m
{ = const}.
The map m is invertible since
m

1 0
Fx Fy

, det(m

) = 1 Fy = 0.
Thus
v(, ) = u(m
1
(, )), u(x, y) = v(m(x, y)).
By this change of variables we derive
ux = v

x + vx
= v

1 + vFx,
uy = v

y + vy
= vFy .
Consequently
0 = aux + buy
= a(v

+ vFx) + bvFy
= av

+ v(aFx + bFy)
= av

+ 0 .
Notice that a = 0 since otherwise the rst integral condition forces a = b = 0 . Hence
v

= 0 v(, ) = f()
for an arbitrary function f of one variable. Actually, this is to be expected since u is
constant along the level curves of F and, as we saw above, the latter are mapped into
the horisontal lines = const . Thus v must depend only on the vertical direction .
Going back to the (x, y) -variables discover
(4.4) u(x, y) = v(m(x, y)) = f(F(x, y)).
(8) Hence solving (4.4) is equivalent to nding the rst integral of the characteristic system
x = a(x, y)
y = b(x, y).
In the the derivation of (4.3) we were saying that " F(x, y) = C was the expression
of the solution to the characteristic system as a level curve", whereas to prove (4.4) we
used the notion of the rst integral of the characteristic system. Let us show that (in
the plane) these objects coincide. That is, nding the rst integral of the plane system
of ODEs is the same as solving it.
On the one hand if F(x, y) = C is the solution formula (in the implicit form) for
the system
x = a(x, y)
y = b(x, y).
Then every solution expressed in the parametric form (x(t), y(t)) must satisfy
F(x(t), y(t)) C,
so F is the rst integral.
On the other hand, suppose that F is the rst integral of the system, so
aFx + bFy = 0.
Write the level curve in a parametric form
F(x, y) = C

x
y

x(t)
y(t)

.
Then dierentiating F(x(t), y(t)) C gives
xFx + yFy = 0.
Consequently comparing the two relations for F we discover that for some function
(t) along the level curve
(t) x = a(x, y)
(t) y = b(x, y).
124A PARTIAL DIFFERENTIAL EQUATIONS 21
Use function
(t) =
Z
(t) dt,

= ,
to reparametrise the level curve as

x
y

x(t)
y(t)

x((t))
y((t))

.
The reparametrised curve satises

x =

x() = a( x, y)

y =

x() = b( x, y).
Thus the level curves of the rst integral when properly parametrised are the solution
curves.
(9) The map m allows to prove the solvability of the nonlinear (more precisely the semi-
linear) equation
a(x, y)ux + b(x, y)uy = H(x, y, u)
following the same strategy. We can prove the existence result for a suitably posed IVP.
To formulate the IVP we argue as follows. The characteristics for the nonlinear
equation are the same as above, they are given by the level curves of F . Suppose the
condition
Fy = 0
holds in a neighbourhood of (0, 0) . The tangent vector to a characteristic is

A =
(Fy, Fx) . Therefore the characteristics are not tangential to the line
{x = 0}
in the neighborhood of (0, 0) . Consider the following problem: for a given smooth
nd u satisfying
(
a(x, y)ux + b(x, y)uy = H(x, y, u) near (0, 0)
u(0, y) = (y) on the line {x = 0}.
We now prove the existence of such u .
The (, ) -variables computation, we have done for (4.4), gives for v(, ) the
equation
av

+ v(aFx + bFy) = H(x, y, v), (x, y) = m


1
(, ).
Introducing
a(, ) = a m
1
(, ),

H(, , v) = H(m
1
(, ), v),
we rewrite the equation as
a(, )v

=

H(, , v) v

= G(, , v),
with
G(, , v) =

H(, , v)
a(, )
.
The neighborhood of (0, 0) is mapped by m onto some neighborhood N of (0,
0
) ,
where
0
= F(0, 0) . The line {x = 0} is mapped onto = 0 . Thus m
1
pulls
back our problem to nding v(, ) satisfying
(
v

= G(, , v) in N
v(0, ) =

() on the line { = 0} N,
with

() = (y)

y=(m
1
(0,))
2
.
This is an IVP for the ODE for v , for which the standard existence-uniqueness theorem
produces a smooth solution v(, ) . Hence we obtain the desired solution
u(x, y) = v(m(x, y)) = v(x, F(x, y)).
(10) Solving problems. For the problem solving the new diculties are
introduced by the right hand side
au
x
+bu
y
= f
and by the lower order terms
au
x
+bu
y
+cu = f.
22 DENIS A. LABUTIN
The right strategy is to use the superposition principle for non homogeneous
equations and integrating factors for the lower order terms.
Problem 4. Find u = u(t, x) solving the initial value problem
2u
t
= u
x
+xu, u(0, x) = 1.
We cannot use any formulas given above due to the term xu . First we
have to get rid of it. Compute
_
xdx =
x
2
2
+C.
The constant is not needed for the integrating factor technique: namely,
multiplying the equation by e
x
2
/2
,
2e
x
2
/2
u
t
= e
x
2
/2
u
x
+xe
x
2
/2
u,
nd that
e
x
2
/2
u
t
=
_
e
x
2
/2
u
_
t
, e
x
2
/2
u
x
+xe
x
2
/2
u =
_
e
x
2
/2
u
_
x
.
Hence we can substitute
v = e
x
2
/2
u
to nd that
2v
t
= v
x
, v(0, x) = e
x
2
/2
.
This is the simplest homogeneous transport equation. We nd at once that
v = f
_
x +
t
2
_
, f(x) = e
x
2
/2
.
Thus
u(t, x) = e
x
2
/2
v
= e
x
2
/2
e
(x+
t
2
)
2
/2
= e
xt
2
+
t
2
8
.
Problem 5. Find u = u(t, x) solving the initial value problem
u
t
+ (1 +t
2
)u
x
+u = 1 u(0, x) = e
x
.
This equation has the variable coecients, the lower order term, and the
nontrivial right hand side. We need to make several steps to reduce it to
the homogeneous equation au
t
+bu
x
= 0 .
First, the easiest way is to combine u
t
+u . The integrating factor can
be taken e
t
. Multiplying deduce
e
t
u
t
+e
t
u +e
t
(1 +t
2
)u
x
= e
t
,
so that
(e
t
u)
t
+ (1 +t
2
)(e
t
u)
x
= e
t
.
Substitute
v = e
t
u
and nd
v
t
+ (1 +t
2
)v
x
= e
t
, v(0, x) = e
x
.
Next, we easily guess that e
t
is solution to the non homogeneous equation.
Therefore
w = v e
t
124A PARTIAL DIFFERENTIAL EQUATIONS 23
satises
w
t
+ (1 +t
2
)w
x
= 0, w(0, x) = e
x
1.
This is a simple homogeneous rst order equation. The characteristics are
dt
1
=
dx
1 +t
2

_
(1 +t
2
) dt =
_
dx t +
t
3
3
+C = x.
Thus the general solution is
w = f
_
x t
t
3
3
_
.
The initial condition gives
f(x + 0) = e
x
1.
Consequently
w = e
x+t+
t
3
3
1,
then
v = e
t
+w = e
t
+e
x+t+
t
3
3
1,
and then
u = e
t
v = 1 +e
x+
t
3
3
e
t
.
(11) Initial value problem for nonlinear transport equations.
5. Initial value problem on the real line
(1) Initial value problem for the wave equation. For the wave equation

c
u = 0
we can nd the general solution
u(t, x) = f(x +ct) +g(x ct),
f, g are arbitrary functions of one variable. This is done using the re-
duction to the transport equations. Then the solution to the initial value
problem on the line
_

c
u = 0, t > 0, x (, +)
u[
t=0
= (x)
u
t
[
t=0
= (x)
can be derived in the form of dAlamberts formula:
u(t, x) =
(x +ct) +(x ct)
2
+
1
2c
_
x+tc
xtc
(s) ds.
(2) The important issue is to visualise in the (t, x) -plane the calculations with
the dAlamberts formula.
24 DENIS A. LABUTIN
(3) Initial value problem for the heat equation. Solutions for the initial
value problem both for the transport and the wave equation were obtained
by the following procedure. First, we derived the formula for the general
solution. Second, we specied the general solution to satisfy the initial
conditions.
Such approach does not work for the heat equation on the line. There
does not exist a simple formula for the general solution to the heat equation.
Instead we have to treat the initial value problem directly.
6. Energy method
(1) In this course the energy method is used only to prove the uniqueness of the
boundary value problems. In fact the energy method is a truly fundamental
tool for analysis of linear and nonlinear PDEs.
(2) If
u
t
= c
2
u
xx
, t > 0, x (a, b),
then the convenient energy is
c[u](t) =
_
b
a
u
2
2
dx =
_
b
a
u(t, x)
2
2
dx.
If
u
tt
= c
2
u
xx
, t > 0, x (a, b),
then the convenient energy is
c[u](t) =
_
b
a
1
2
_
u
2
t
+c
2
u
2
x
_
dx =
_
b
a
1
2
_
u
t
(t, x)
2
+c
2
u
x
(t, x)
2
_
dx.
(3) The energy method computations always involve evaluating of
d
dt
c[u](t)
by dierentiation under the integral sign, using the equation for replacing
some derivatives by the other, integration by parts relying on the given
boundary conditions.
7. Reflection method for problems on the half line
(1) Method of reections allows to derive the solutions to some (but not all!)
problems with the boundary conditions (some, not all!) from the solutions
to the initial value problem on (, +) . Thus the reection method
can be summarized as the reduction
I.V.P. for Lu = 0 on x (0, +)
with boundary conditions at x = 0

I.V.P. for Lu = 0 on x (, +)
without the boundary
Below we introduce the reections approach through some examples.
124A PARTIAL DIFFERENTIAL EQUATIONS 25
(2) Problems with homogeneous boundary conditions. Consider the
wave equation on the half-line with the zero Dirichlet boundary condition:
_

_
u =0, t > 0, x (0, +)
u[
t=0
=(x), x > 0
u
t
[
t=0
=(x), x > 0
u[
x=0
=0, t > 0.
Here is the argument allowing to write down the solution to the problem
using the dAlamberts formula. Take the odd reection of , from
(0, +) to (, 0) :

,

(x) =
_

_
, (x) for x > 0
0 for x = 0
, (x) for x < 0.
Consider the initial value problem on (, +) with the reected data

,

:
_

_
v = 0, t > 0, x (, +)
v[
t=0
=

(x)
v
t
[
t=0
=

(x).
The data

,

are the odd functions. Hence by the property of the wave
equation the solution v is odd in x for any t . But an odd function must
be equal to zero at the origin, thus
v(t, 0) = 0 for all t.
Hence v for x 0 is the solution to the problem for u since it satises
both the initial and the boundary conditions there. Using the dAlamberts
formula
u(t, x) =

(x +t) +

(x t)
2
+
1
2
_
x+t
xt

(s) ds .
Using the visualisation of the dAlamberts formula in the (t, x) -plane and
the oddness of the data it is easy to express the answer using only ,
(not

,

). The formula will be somewhat lengthy but it admits a simple
visualisation on the (t, x) -plane.
(3) The method of reection allows to solve some boundary value problems on
the nite interval (0, ) . For example, one can solve the Dirichlet problem
for the wave equation
_

_
u = 0, t > 0, x (0, )
u[
t=0
= (x), x (0, )
u
t
[
t=0
= (x), x (0, )
u[
x=0
= u[
x=
= 0, t > 0.
The solution will have an incredibly lengthy formal description, but it has
a clean visualisation on the (t, x) -plane.
(4) The homogeneous Neumann boundary conditions are also manageable by
the reection method. For example, consider the heat equation on the
26 DENIS A. LABUTIN
half-line
_

_
u
t
u
xx
= 0, t > 0, x (0, +)
u[
t=0
= (x), x > 0
u
x
[
x=0
= 0, t > 0.
Take the even reection of from (0, +) to (, 0) :

(x) =
_
(x) for x 0
(x) for x 0.
Consider the initial value problem on (, +) with the reected datum

:
_
v
t
v
xx
= 0, t > 0, x (, +)
v[
t=0
=

(x) .
The datum

is an even function. Hence by the property of the heat
equation the solution v is even in x for any t . But an even (smooth)
function must have zero derivative at the origin, thus
v
x
(t, 0) = 0 for all t.
Hence v is the solution to the problem for u since satises both the initial
and the boundary conditions there. Using the heat kernel
u(t, x) =
_
+

S(t, x y)

(y) dy .
Using the evenness property of the heat kernel and the evenness of

it is
easy to express the answer using alone (instead of

).
(5) Reections for the non-homogeneous boundary conditions.
Method of reections allows also to reduce some problems with the nonho-
mogeneous boundary conditions to the IVP on (, +) . For example,
consider the general Dirichlet problem for the wave equation on (0, +) :
_

_
u
tt
c
2
u
xx
= 0, t > 0, x (0, +)
u[
t=0
= (x), x > 0
u
t
[
t=0
= (x), x > 0
u[
x=0
= h(t), t > 0.
The Dirichlet boundary condition is now given by h(t) , h ,= 0 . Here is
the reduction argument. Recall that

c
f(x ct) = 0
for an arbitrary function f of one variable. Choose f() = h( /c) , so
that
f(x ct) = h
_
t
x
c
_
.
Dene
v(t, x) = u(t, x) h
_
t
x
c
_
,
so that v(t, 0) = h(t) h(t) = 0 for all t > 0 . Then v solves
_

_
v
tt
c
2
v
xx
= 0, t > 0, x (0, +)
v[
t=0
=
1
(x), x > 0
v
t
[
t=0
=
1
(x), x > 0
v[
x=0
= 0, t > 0,
124A PARTIAL DIFFERENTIAL EQUATIONS 27
with
1
(x) = (x) h(x/c) ,
1
(x) = (x) h

(x/c) . This is the


problem with the zero Dirichlet condition. We solved it above using the
odd reection. Hence the answer is
for x > ct :
u(t, x) =

1
(x +ct) +
1
(x ct)
2
+
1
2c
_
x+tc
xtc

1
(s) ds
+h
_
t
x
c
_
=
(x +ct) +(x ct)
2
+
1
2c
_
x+tc
xtc
(s) ds

h((x +ct)/c) +h((x ct)/c)


2

1
2c
_
x+tc
xtc
h

(s/c) ds
+h
_
t
x
c
_
=
(x +ct) +(x ct)
2
+
1
2c
_
x+tc
xtc
(s) ds

h((x +ct)/c) +h((x ct)/c)


2
+
1
2
_
(x+tc)/c
(xtc)/c
h

(r) dr
+h
_
t
x
c
_
=
(x +ct) +(x ct)
2
+
1
2c
_
x+tc
xtc
(s) ds
h((x ct)/c)
+h
_
t
x
c
_
=
(x +ct) +(x ct)
2
+
1
2c
_
x+tc
xtc
(s) ds;
for ct x > 0 using the similar computations:
u(t, x) =
(x +ct) +(x ct)
2
+
1
2c
_
x+tc
xtc
(s) ds
+h
_
t
x
c
_
.
The meaning of the solutions is simple. For x > ct the wave launched by
the motion h(t) at the origin does not reach the point x , and the solution
is given by the usual dAlamberts formula. For ct > x the solution is the
superposition of the free wave coming from the dAlambert and the wave
launched from the origin by the prescribed motion h(t) .
28 DENIS A. LABUTIN
8. Fundamental solution and Duhamels principle
(1) Delta function. Delta function is not a function! The meaning will
be assigned not to itself, but to expressions involving it. Therefore for
us formulas as
_
+

(x y) dy = 1, (x) = 0 x ,= 0, . . .
will have a meaning, but itself will not. The formulas involving -
function (such as above) express the limiting results of the corresponding
processes.
(2) Intuitively -function expresses such physical notions as the density of the
point electrostatic charge, the energy of the point heat source, and similar.
This is done using the limit of -shaped functions in the same way as in
physics one deals with the point-concentrated objects as limits of smeared
ones, as smearing gets smaller and smaller.
A -shaped function

(x) is the function satisfying:


(a)
_
+

(x) dx = 1 ;
(b)

(x) 0 for all x ;


(c)

(x) = 0 for all x with [x[ > (or lim


0

(x) = 0 for every


x ,= 0 .)
The following two examples of -shaped sequences are of fundamental
importance for this course.
First, the characteristic function approximation of is
(8.5)
1

(x) =
_
0 , x / (0, )
1/ , x [0, ],
where

(x) =
_
0 , x / (0, )
1 , x [0, ].
Second, the heat kernel approximation is
1

G(x/

) =
1

4
e

x
2
4
.
Now, -function is the "limit" of a -shaped sequence as 0 . Thus
(8.6)

(x),
1

(x),
1

4
e

x
2
4
(x), 0.
The limit in (8.6) is understood in the following sense: the validity of any
given formula
F() = f
means that after the substitution F(

) and letting 0 we nd
lim
0
F(

) = f.
(3) Let us illustrate the technique of dealing with .
Problem 6. Let us show that
_
+

(x) dx = 1.
124A PARTIAL DIFFERENTIAL EQUATIONS 29
Indeed, take any -shaped sequence

(for example

/ ). We have
_
+

(x) dx = 1 > 0,
hence trivially
lim
0
__
+

(x) dx
_
= 1.
Thus according to our denition we may write
_
+

(x) dx = 1.
Problem 7. Show that
(x) = 0 x ,= 0.
Problem 8. Show that for any continuous function f
_
+

f(x) (x) dx = f(0).


To prove this formula it is convenient to use the -shaped sequence

(x) .
Problem 9. Fix x
0
. For the shifted (x x
0
) show that
(x x
0
) = 0 x ,= x
0
,
_
+

(x x
0
) dx = 1,
and
_
+

f(x) (x x
0
) dx = f(x
0
).
Problem 10. Prove that for a continuous function f
(8.7)
_
+

f(y) (x
0
y) dy = f(x
0
).
The formula can be approximately rewritten as
f(x
0
)
_
+

f(y)

(x
0
y) dy

(x
0
y
j
) f(y
j
) y
j

j
(x
0
y
j
) f(y
j
) y
j
=

j
f(y
j
)y
j
(x y
j
)

x=x0
.
This form emphasizes the important meaning of (8.7). Namely, an arbitrary
force (or signal, or eld, or ) f(x) at a point x is approximately equal
to the value at x of the sum of the point-impulse forces (or signals, or
elds, or ) shifted to y
j
:
(8.8) f(x)

j
f(y
j
) (y
j+1
y
j
) (x y
j
).
30 DENIS A. LABUTIN
To derive (8.7) just use
_
+

f(y)

(x
0
y) dy =
_

+
f(x
0
z)

(z) (dz) since x


0
y = z, dy = dz
=
_
+

f(x
0
z)

(z) dz
f(x
0
z)

z=0
.
Problem 11. For the Heaviside step-function
(x) = H(x) =
_
0, x < 0
1, x 0
show that
(x)

= (x).
For that smear the step-function

(x) on scale . Then show that

(x) =

(x)

is a -shaped sequence.
(4) Fundamental solution. Consider the initial value problem for a very
general evolution equation with the unknown u(t, x)
_

t
u = Lu t > 0, x (, +)
u[
t=0
= (x).
Think that
L = L(
x
)
is, say, a constant coecient operator involving only
x
,
2
x
, . . . . Many
important problems from sciences have such form. For example, we have
already encountered the heat diusion L = k
2

2
x
, and transport L =
c
x
.
(5) The fundamental solution (or propagator) is the evolution of the initial
point-impulse. That is, the solution of
_

t
u = Lu, t > 0, x (, +)
u[
t=0
= (x)
This means that

t
K(t, x) L(
x
)K(t, x) = 0, t > 0, x (, +)
and K(t, x) is -shaped in the same sense as in (8.6) with relaced by
t . That is
K(t, x) (x), t 0.
Hence the fundamental solution shows how the equation smears for t > 0
the initial (at time t = 0 ) datum .
As we will see below, knowing the propagator for L is the key for solv-
ing the initial value problem with arbitrary data. This explains the name
"fundamental". We will denote the propagator by K(t, x) or K
L
(t, x) .
(6) For example, the Gaussian (heat kernel)
G(t, x) =
1

4t
e

x
2
4t
satises
_

2
x
_
G(t, x) = 0, t > 0.
Moreover, it forms a -shaped sequence,
G(t, x) (x), t 0.
124A PARTIAL DIFFERENTIAL EQUATIONS 31
Hence the Gaussian is the fundamental solution to the heat equation.
(7) Somewhat unusual example is the fundamental solution to the transport equation
tu + cxu = 0.
One easily makes the formal computations for K(t, x) = (x ct) :
x(x ct) =

xct
1,
t(x ct) =

xct
(c),
so that
(t + cx)(x ct) = c + c = 0.
We also have (x ct) (x) as t 0 (without any smearing). Hence (x ct) is
the fundamental solution of the transport equation.
This answer has a clear physical meaning. It says that the initial point-mass with
the density is transported by the one dimensional ow with the velocity c along
the x -access without smearing or spreading. That is indeed what the perfect transport
(without difusion) does!
(8) We veried that two suggested functions are the fundamental solutions of
the heat and transport equation. There is a general method based on the
Fourier transform for the computation of the fundamental solution for a
given equation from the scratch. It will be studied in the second quarter.
(9) Fundamental solution and the formula for IVP. Let us show that the
superposition principle implies that the solution of
_

t
u = Lu, t > 0, x (, +)
u[
t=0
= (x)
is given by
(8.9) u(t, x) =
_
+

K(t, x y)(y) dy.


Hence, knowing the fundamental solution K(t, x) allows us to calculate
the solution with any initial data by the simple integration!
Verifying that (8.9) solves the IVP is easy. Indeed, rstly for t > 0 and
x (, +) we have
(
t
L(
x
))
__
+

K(t, x y)(y) dy
_
=
_
+

(
t
L(
x
))K(t, x y)(y) dy
=
_
+

(
s
K(s, z) L(
z
)K(s, z))

(s,z)=(t,xy)
(y) dy
=
_
+

0(y) dy
= 0.
Secondly, for every x (, +)
lim
t0
__
+

K(t, x y)(y) dy
_
=
_
+

(x y)(y) dy
= (x).
Instead of verifying the formula given by somebody, one can derive it as
shown below.
(10) The basic building block for the derivation of (8.9) is the following simple observation.
Consider the IVP with the shifted :
C(x a).
32 DENIS A. LABUTIN
That is
(
tu = Lu, t > 0, x (, +)
u|
t=0
= C(x a).
Use the linearity to prove that the solution is
u = CK(t, x a),
where K is the fundamental solution. This is the formula we will need.
(11) Now x a small > 0 , discretize the x variable
x
j
= j, j = 0, 1, 2 . . . ,
and approximate our datum
(x)
X
j
(x
j
)(x x
j
).
The key observation is to write
(x
j
)(x x
j
) = (x
j
)
1

(x x
j
)
with the characteristic function approximation of from (8.5). Therefore
(x
j
)
1

(x x
j
) (x
j
) (x x
j
).
The summation then gives
(8.10) (x)
X
j
(x
j
) (x x
j
),
which is nothing but the physical approximation of a general charge distribution by a
sum of point charges. That is we approximated by a sum of discrete impulses located
at points x
j
.
Notice that we could skip the details of the discretization and at once deduce (8.10)
from (8.8)!
(12) Next, by the linearity
u(t, x)
X
j
u
j
(t, x),
where u
j
solves
(
tu
j
= Lu
j
, t > 0, x (, +)
u
j
|
t=0
= (x
j
) (x x
j
).
By our building block solution we already know that
u
j
(t, x) = (x
j
)K(t, x x
j
).
Consequently
u(t, x)
X
j
u
j
(t, x)
=
X
j
(x
j
)K(t, x x
j
)
=
X
j
K(t, x x
j
)(x
j
)(x
j+1
x
j
).
Finally, notice that the last sum approximates an integral over x . Letting the discreti-
sation
x
j+1
x
j
= 0
we discover
u(t, x) = lim
0
0
@
X
j
K(t, x x
j
)(x
j
)(x
j+1
x
j
)
1
A
=
Z
+

K(t, x y)(y) dy,


which is desired (8.9).
124A PARTIAL DIFFERENTIAL EQUATIONS 33
(13) To make the formulae shorter, it is convenient to introduce the solution
operator e
tL
acting on the functions F on R . Dene
(e
tL
F)(x) =
_
+

K(t, x y)F(y) dy.


In the case L is a number e
tL
is the usual exponential. In these notations
we have
_

t
u = Lu, t > 0, x R
u[
t=0
= (x)
u(t, x) = e
tL
(x)
(14) The propagator approach gives the same formula for any constant coe-
cients linear operator L . Notice, however, that the wave equation does
not t immediately in this form we have
2
t
instead of
t
in the initial
value problem for the wave equation. Later we will show how to adapt the
arguments for the wave equation.
(15) Duhamels principle. Knowing the propagator also allows us to solve
the non-homogeneous problem
(8.11)
_

t
u =Lu +f(t, x), t > 0, x R
u[
t=0
=0
via the Duhamels formula:
u(t, x) =
_
t
0
e
(ts)L
f(s) ds
=
_
t
0
_
+

K(t s, x y)f(s, y) dy ds.


Again, once the formula is given, its verication is not hard. Firstly,
u(0, x) =
_
0
0
_

_
dt = 0,
therefore the initial condition holds. Secondly, equality
L(
x
)u(t, x) =
_
t
0
_
+

L(
x
)K(t s, x y)f(s, y) dy ds
=
_
t
0
_
+

L(
z
)K(r, z)

(r,z)=(ts,xy)
f(s, y) dy ds
together with

t
u(t, x) =
_
+

K(t s, x y)f(s, y) dy

s=t
+
_
t
0
_
+

t
K(t s, x y)f(s, y) dy ds
=
_
+

(x y)f(t, y) dy
+
_
t
0
_
+

r
K(r, z)

(r,z)=(ts,xy)
f(s, y) dy ds
34 DENIS A. LABUTIN
imply
(
t
L)u =
_
+

(x y)f(t, y) dy
+
_
t
0
_
+

r
K(r, z) L(
z
)K(r, z)
_

(r,z)=(ts,xy)
f(s, y) dy ds
=f(t, x) +
_
t
0
_
+

0 f(s, y) dy ds
=f(t, x).
(16) We now derive Duhamels formula on the "physical" level of rigor. For that we shall use
the following basic building block problem. Consider the delayed IVP:
(
tu = Lu, t > s, x (, +)
u|t=s = (x).
Using the invariance in time, one can easily see that the value of the solution to the
delayed problem is
u(t, x) =
Z
+

K(t s, x y)(y) dy
where K is the fundamental solution. This is the formula we will need.
(17) Fix T > 0 . To compute u(T, x) solving (8.11) we rst split the time interval [0, T]
into N segments of the length t = T/N ,
0 = t
0
< t
1
= t
< t
2
= 2t = t
1
+ t

< t
j
= jt = t
j1
+ t

< t
N
= T.
Next, in (8.11) discretise f in time by splitting it into
f(t, x)
N1
X
j=0
f
j
(t, x), f
j
(t, x) = f(t
j+1
, x)
(t
j
,t
j
+t)
(t).
By the linearity
u(T, x)
N1
X
j=0
u
j
(T, x),
where u
j
(t, x) solves
(
tu
j
= Lu
j
+ f
j
, t > 0, x (, +)
u
j
|
t=0
= 0.
The point here is that the right hand side f
j
(t, x) is identical zero if 0 < t < t
j
and
t
j
+t < t < T . Only on the short time interval (t
j
, t
j
+t) it acts on the equation.
Let us evaluate u
j
(T, x) for every j . By (8.9) we have
u
j
(t, x) = 0 for 0 < t t
j
.
Ignoring the tiny time interval (t
j
, t
j+1
) of length t we use the rst order approxi-
mation to evaluate for t = t
j+1
u
j
(t
j+1
, x) u
j
(t
j
, x) + t tu
j
(t
j
, x)
= 0 + t

L(x)u
j
+ f
j

(t
j
,x)
= t

0 + f(t
j+1
, x)

.
Consequently for t
j+1
< t < T our u
j
solves
(
tv = Lv, t
j+1
< t < T, x (, +)
v|t=t
j+1
= f(t
j+1
, x)t.
124A PARTIAL DIFFERENTIAL EQUATIONS 35
But this is just the delayed IVP. Hence, as we shown above
u
j
(T, x) =
Z
+

K(T t
j+1
, x y)f(t
j+1
, y)t dy.
Therefore
u(T, x)
N1
X
j=0
Z
+

K(T t
j+1
, x y)f(t
j+1
, y) dy

t.
Recognising the Riemann integral sum here we discover letting t 0 that
u(T, x) =
Z
T
0
Z
+

K(T s, x y)f(s, y) dy ds.


(18) Notice that the Duhamels formula gives the following statement for the
heat equation. If (x) is odd (even) and for every t the function f(t, x)
is odd (even) in x , then the solution of
_

t
u = Lu +f(t, x), t > 0, x (, +)
u[
t=0
= (x),
u = u(t, x) is odd (even) in x for any xed t .
This follows at once from the representation
u(t, x) =
_
+

K(t, x y)(y) dy +
_
t
0
_
+

K(t s, x y)f(s, y) dy ds
with
K(t, z) =
1

t
G
_
z

t
_
,
and the evenness of the Gaussian
G(x) =
1

4
e

x
2
4
.
9. Boundary value problems and separation of variables
(1) Separation of variables method for the heat and wave equations with either
Dirichlet or Neumann boundary conditions is familiar from calculus. Actu-
ally the method works for more general equations and boundary conditions.
(2) Suppose f is a given function on a nite segment [a, b] .
Boundary conditions

1
f(a) +
1
f(b) +
1
f

(a) +
1
f

(b) = 0

2
f(a) +
2
f(b) +
2
f

(a) +
2
f

(b) = 0
are called symmetric if for any functions f , g satisfying
them one has
f

(x)g(x) f(x)g

(x)

x=b
x=a
= 0.
This denition is not very inspiring. The true meaning of the symmetry
is the symmetry of the operator d
2
/dx
2
with respect to the L
2
inner
product. Indeed,
36 DENIS A. LABUTIN
the boundary conditions are symmetric if and only for any smooth
functions f , g satisfying them one has
_
d
2
f
dx
2
, g
_
L
2
=
_
f,
d
2
g
dx
2
_
L
2
(3) The theoretical basis for the separation of variables method is provided by
the following theorem. Its most important part (iii) states that the sepa-
ration of variables method always works provided the boundary conditions
are symmetric.
Theorem. Consider the eigenvalue problem
d
2
dx
2
f = f on (a, b)
with some symmetric boundary conditions. Then:
(i) all eigenvalues are real and form an innite discrete sequence

1

2
,

n
+ as n ;
(ii) the corresponding eigenfunctions are L
2
-orthogonal (f
n
, f
m
)
L
2 = 0 ;
(iii) any L
2
can be expanded as
=

n=1
A
n
f
n
on [a, b]
with the coecients given by
A
n
=
(, f
n
)
L
2
(f
n
, f
n
)
L
2
.
The convergence of the series in the theorem is in the L
2
sense. The
condition
n
implies that the multiplicity of every eigenvalue is
nite. The theorem unites several theorems in Strauss.
(4) We know from the calculus that any can be expanded either into the
full Fourier series, or into the cos -half range Fourier series, or into the
sin -half range series. The theorem provides us with the expansion with
respect to the general system f
n
of the eigenfunctions not covered by
the three calculus cases.
For example, the mixed boundary conditions
u[
x=0
= 0, u
x
[
x=l
= 0
on (0, l) lead through the separation to the eigenvalue problem
f

= f, f(0) = f

(l) = 0.
The eigenvalues turn out to be

n
=
_
(n +
1
2
)
l
_2
, n = 0, 1, 2, . . . ,
and the eigenfunctions
f
n
(x) = sin
_
(n +
1
2
)x
l
_
.
This is neither the full Fourier nor a half range Fourier system. Hence the
calculus expansions do not work in this case. However, the theorem tells
124A PARTIAL DIFFERENTIAL EQUATIONS 37
that any initial datum on [0, l] can be expanded with respect to f
n
.
Indeed, by the theorem
(x) =

n=0
A
n
sin
_
(n +
1
2
)x
l
_
, x [0, l],
with the coecients computed by the formula
A
n
=
_
l
0
(x) sin
_
(n +
1
2
)x
l
_
dx
_
l
0
sin
_
(n +
1
2
)x
l
_
sin
_
(n +
1
2
)x
l
_
dx
.

Вам также может понравиться