Professional Documents
Culture Documents
MIT18 152F11 Lec 15
MIT18 152F11 Lec 15
Type
Elliptic
Well-posed problems
Boundary value problems: All of Rn (with
boundary conditions
at ); finite boundaries under Dirichlet,
Neumann,
Robin,
or Mixed boundary
conditions
t u(t, x) u(t, x) = f (t, x)
Diffusive (parabolic): Initial value (Cauchy)
problems: all of Rn
at t = 0; Initial +
boundary value problems: data at t = 0 +
Dirichlet, Neumann,
Robin,
or Mixed
boundary conditions
t2 u(t, x) + u(t, x) = f (t, x) Hyperbolic
Initial value (Cauchy)
problems: all of Rn
at t = 0; Initial +
boundary value problems: data at t = 0 +
Dirichlet, Neumann,
Robin,
or Mixed
boundary conditions
Features
mean value properties;
maximum principle; Harnack inequality
2. Motivating example
Lets consider the following second-order linear PDE on R1+n :
(2.0.1)
def
Lu = A u + B u + Cu = 0.
In (2.0.1), A, B, C are allowed to be functions of the coordinates (x0 , , xn ). We will also use the
standard notation x0 = t. By the symmetry of the mixed partial derivatives, we can also assume
that A is symmetric:
1
A = A .
(2.0.2)
The question we would like to address at the moment is the following: what are the basic properties
of solutions to (2.0.1)? Is this equation most like a Laplace, heat, or wave equation? That is, is
(2.0.1) elliptic, diffusive, or hyperbolic? As we will see, the most important part of equation (2.0.1)
in this context is the principal part A u, which involves the top-order derivatives.
To begin answering this question, lets start with a simple example on R2 . Lets try to classify
the following equation:
(2.0.3)
def
Lu = t2 u 4t x u + 2x2 u = 0.
Note that it would be easy to answer our question if we were able to make a linear change of
variables that eliminates the cross term 4t x u; the PDE would then look just like one of the
other ones we have already studied. More precisely, lets try to eliminate the cross terms by making
good choices for the constants a, b, c, d in the following linear change variables:
(2.0.4a)
e
t = at + bx,
(2.0.4b)
x
e = ct + dx.
In order to have a viable change of variables, we also need to achieve the following non-degeneracy
condition from linear algebra:
ad bc 6= 0.
(2.0.5)
(2.0.5) states the determinant of the above linear transformation is non-zero, and that the transformation is non-degenerate.
Then using the chain rule, we have that
e
te
x
et +
xe = aet + cxe,
t
t
te
x
e
xe = bet + dxe.
(2.0.6b)
x =
et +
x
x
Inserting (2.0.6a) - (2.0.6b) into (2.0.3), we compute that
(2.0.6a)
(2.0.7)
t =
Lu = (a2 4ab + 2b2 )et2 u + (2ac + 4bd 4ad 4bc)et xeu + (c2 4cd + 2d2 )xe2 u.
a = 1,
b = 0,
c = 2,
d = 1.
Note that (2.0.8) also verifies the non-degeneracy condition (2.0.5). We remark that other choices
would also have worked. In the new coordinates, we have that
(2.0.9)
Lu = et2 u 2xe2 u.
Dividing by 2, we see that the PDE (2.0.3) was actually a standard linear wave equation in
disguise:
1
et2 u + xe2 = 0.
2
e
Relative to the coordinates (t, x)
e , the speed associated to the wave equation (2.0.10) is 2.
Lets do another example. Consider the PDE
(2.0.10)
def
Lu = 2t2 u 2t x u x2 u + x u = 0.
(2.0.11)
Lu = (2a2 2ab b2 )et2 u + (2ac bd 2ad 2bc)et xeu + (2c2 4cd d2 )xe2 u
+ bet u + dxeu.
Choosing
(2.0.13)
1
a= ,
2
c = 1,
b = 0,
d = 1,
we see that
Lu = et2 u xe2 u + xeu.
(2.0.14)
Thus, multiplying by 1, we see that (2.0.11) is really just a Laplace-like equation in disguise:
et2 u + xe2 u xeu = 0.
(2.0.15)
Equation (2.0.11) is therefore elliptic. We remark that the first-order term in (2.0.15) does not
affect the elliptic nature of the system.
Lets do one final example. Consider the PDE
def
Lu = t2 u 2t x u + x2 u + x u = 0.
(2.0.16)
Lu = (a2 2ab + b2 )et2 u + (2ac + 2bd 2ad 2bc)et xeu + (c2 2cd + d2 )xe2 u
+ bet u + dxeu.
Choosing
(2.0.18)
a = 1,
b = 0,
c = 1,
we see that
(2.0.19)
Thus, (2.0.16) is equivalent to
Lu = et2 u xeu.
d = 1,
xeu + et2 u = 0.
(2.0.20)
Now observe that (2.0.20) is just the standard heat equation, with the variable x
e playing the role
of time and e
t playing the role of space. Equation (2.0.20) is therefore diffusive (parabolic).
3. A general framework
In this section, we will establish a general framework for classifying second order constant coefficient scalar PDEs. The framework will cover the three examples from the previous section as
special cases. The proof will reveal that the classification is intimately connected to the theory of
quadratic forms from linear algebra. Throughout this section, we will use the notation
x = (x0 , x1 , , xn ).
(3.0.21)
Lu = A u + B u + Cu = 0,
(3.0.22)
where A = A .
We begin by providing a simple version of Hadamards classic definitions.
Definition 3.0.1 (Hadamards classification of second order scalar PDEs). Equation (3.0.22)
is respectively said to be elliptic, hyperbolic, or parabolic according to the following conditions on
the (1 + n) (1 + n) symmetric matrix A :
All of the eigenvalues of A have the same sign - elliptic
n of the eigenvalues of A have the same (non-zero) sign, and the remaining one has the
opposite (non-zero) sign - hyperbolic
n of the eigenvalues of A have the same (non-zero) sign, and the remaining one is 0 parabolic
Remark 3.0.1. Many of the ideas in this section, including the definition above, can be generalized
to include the case where A depends on (x), or even on the solution u itself; PDEs of the latter
type are said to be quasilinear.
We now state and prove the main classification theorem.
Theorem 3.1 (Classification of second order constant-coefficient PDEs). Consider the
following second order constant coefficient PDE
def
(3.0.23)
def
Cu(y) = 0, where y = (m1 ) y y is the standard linear wave operator, and (m)1 =
diag(1, 1, 1, , 1) is the standard Minkowskian matrix.
If n eigenvalues (1) , , (n) of A have the same (non-zero) sign, and the remaining one is
2
0 1
n
e 0 0 u(y 0 , y 1 , , y n )+Pn
(0) = 0, then (3.0.23) can be written as Lu = B
i=1 (y i )2 u(y , y , , y )+
y
Pn e i
0 1
n
(0) (1)
(n)
be a corresponding
i=1 B y i u(y , y , , y ) + Cy = 0. Furthermore, let v , v , , v
P
()
diagonalizing unit-length co-vector basis. More precisely, this means that n=0 |v |2 = 1
() ()
() ()
for 0 n, that A v v = () if = , and that A v v = 0 if =
6 (standard
linear algebraic theory guarantees the existence of such a basis). Then if the non-zero vector
(0)
e0 =
B satisfies B v =
6 0, we also have that B
6 0.
Remark 3.0.2. The sign above distinguishes whether or not most of the eigenvalue of A
are positive or negative. For example, if all of the eigenvalues of A are positive, then Lu =
y u(y) + , while if they are all negative, then Lu = y u(y) + (and similarly for the other
two cases).
Proof. Lets consider the first case, in which all of the eigenvalues have the same (non-zero) sign.
Then by standard linear algebra, since A is symmetric and positive definite (perhaps after multiplying it by 1), there exists an invertible change-of-basis matrix M such that
M A M = I ,
(3.0.24)
def
(3.0.25)
v()
()|
|
(no summation in ),
P
()
()
where () is the eigenvalue of A corresponding to the unit-length covector v (i.e., n=0 |v |2 =
1) appearing in the statement of the theorem.
y
We now make the linear change of variables y = M x . Then by the chain rule, x = x
y =
M y . Therefore,
A
(3.0.26)
u
=
A
M
M
u
=
I
u = y u.
x x
y y
y y
(3.0.27)
def
u = A M M u = (m1 ) u = y u.
x x
y y
y y
In the third case, in which n of the eigenvalues of A have the same (non-zero) sign, and the
remaining one is 0, the proof is similar. The key difference is that because of the zero eigenvalue,
(3.0.24) is replaced with
M A M = D ,
(3.0.29)
def
(3.0.30)
X 2
u
=
D
u
=
u.
u
=
A
M
M
x x
y y
y y
(y i )2
i=1
u = M B u.
x
y
Thus, using using (3.0.25), we have that
(3.0.31)
(3.0.32)
e 0 def
B
= M0 B = v(0) B 6= 0.
=
1 2
2 2
det(A I) = det
1 2
2 2
= 2 3 2 = 0.
17
(3.0.35)
=
.
2
Since the eigenvalues are of opposite sign, the corresponding PDE is hyperbolic.
3
=
2 1
1 1
.
det(A I) = det
2
1
1
1
= 2 + 3 + 1 = 0.
3 5
(3.0.38)
=
.
2
Both of these eigenvalues are negative, and thus the corresponding PDE is elliptic.
Example 3.0.3. In the final example from above,
(3.0.39)
=
1 1
1 1
.
det(A I) = det
1 1
1 1
= 0, 2,
= 2 + 2 = 0.
MIT OpenCourseWare
http://ocw.mit.edu
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.