38

3.1 Classification and Standard Form Reduction

 

2
2 u
u 2
2 u 2 u 2
+2
+
+
,
+ 2
2
x
x x
x
x
x2
 
 
2u
2 u 2
2 u 2 u 2 u 2
u 2
+ 2
= 2
+2
+
+
,
2
y

y
y y

y
y 2
y 2


2u
2 u u 2
2 u
2 u

u 2
+ 2
=
+
+
+
+
.
xy
2 x y x y y x
x y
xy
xy
2u
2u
= 2
2
x

Chapter 3

The equation becomes

Second Order Linear and

Semilinear Equations in Two
Variables

3.1

Classification and Standard Form Reduction . . . . . . . . . . . .

Extensions of the Theory . . . . . . . . . . . . . . . . . . . . . . . .

37
44

We write explicitly only the principal part of the PDE, involving the highest-order derivatives
of u (terms of second order).
It is easy to verify that
(B 2 AC) = (b2 ac)

Consider a general second order linear equation in two independent variables

2u
2u
u
2u
u
+ c(x, y) 2 + d(x, y)
+ e(x, y)
+ f (x, y)u = g(x, y);
+ 2b(x, y)
x2
xy
y
x
y

in the case of a semilinear equation, the coefficients d, e, f and g could be functions of x u, y u

and u as well.
Recall, for a first order linear and semilinear equation, a u/x+b u/y = c, we could define
new independent variables, (x, y) and (x, y) with J = (, )/(x, y) 6= {0, }, to reduce
the equation to the simpler form, u/ = (, ).
For the second order equation, can we also transform the variables from (x, y) to (, ) to put
the equation into a simpler form?
So, consider the coordinate transform (x, y) (, ) where and are such that the Jacobian,
 


(, )
 x y 
J=
=
 6= {0, }.
(x, y)  
x

Then by inverse theorem there is an open neighbourhood of (x, y) and another neighbourhood
of (, ) such that the transformation is invertible and one-to-one on these neighbourhoods.
As before we compute chain rule derivations
u
u
u
=
+
,
x
x x

u
u
u
=
+
,
y
y
y
37

 2

2

+ 2b
,
+c
x
x y
y






B=a
+b
+
,
+c
x x
x y y x
y y
 2
 2

+ 2b
.
C=a
+c
x
x y
y
A=a

Classification and Standard Form Reduction

a(x, y)

(3.1)

where

Contents
3.1
3.2

2u
2u
2u
+ 2B
+C
+ F (u , u , u, , ) = 0,
2

x y y x

2

where (x y y x )2 is just the Jacobian squared. So, provided J 6= 0 we see that the
sign of the discriminant b2 ac is invariant under coordinate transformations. We can use
this invariance properties to classify the equation.
Equation (3.1) can be simplified if we can choose and so that some of the coefficients A,
B or C are zero. Let us define,
D =

/x
/y

and D =

/x
;
/y

then we can write

A = aD2 + 2bD + c

2

B = (aD D + b (D + D ) + c)
,
y y
 2

.
C = aD2 + 2bD + c
y

Now consider the quadratic equation

aD 2 + 2bD + c = 0,

(3.2)

Chapter 3 Second Order Linear and Semilinear Equations in Two Variables

whose solution is given by
D=

b2

ac

39

If the discriminant b2 ac 6= 0, equation (3.2) has two distinct roots; so, we can make both
coefficients A and C zero if we arbitrarily take the root with the negative sign for D and
the one with the positive sign for D ,

/x
b b2 ac
D =
=
A = 0,
(3.3)
/y
a

2
/x
b + b ac
D =
=
C = 0.
/y
a
Then, using D D = c/a and D + D = 2b/a we have
B=

2
ac b2
B 6= 0.
a
y y

Furthermore, if the discriminant b2 ac > 0 then D and D as well as and are real. So,
we can define two families of one-parameter characteristics of the PDE as the curves described
by the equation (x, y) = constant and the equation (x, y) = constant. Differentiate along
the characteristic curves given by = constant,
d =

3.1 Classification and Standard Form Reduction

1. If b2 > ac we can apply the change of variable (x, y) (, ) to transform the original
PDE to
2u
+ (lower order terms) = 0.

In this case the equation is said to be hyperbolic and has two families of characteristics
given by equation (3.4) and equation (3.5).
2. If b2 = ac, a suitable choice for still simplifies the PDE, but now we can choose
arbitrarily provided and are independent and the equation reduces to the
form
2u
+ (lower order terms) = 0.
2
The equation is said to be parabolic and has only one family of characteristics given by
equation (3.6).
3. If b2 < ac we can again apply the change of variables (x, y) (, ) to simplify the
equation but now this functions will be complex conjugate. To keep the transformation
real, we apply a further change of variables (, ) (, ) via
= + = 2 <(),

= i( ) = 2 =(),
2u
2u
2u
=
+

2 2
so, the equation can be reduced to

dx +
dy = 0,
x
y

and make use of (3.3) to find that this characteristics satisfy

dy
b + b2 ac
=
.
dx
a

i.e.,

If the discriminant b2 ac = 0, equation (3.2) has one unique root and if we take this root
for D say, we can make the coefficient A zero,

In this case the equation is said to be elliptic and has no real characteristics.
The above forms are called the canonical (or standard) forms of the second order linear or
semilinear equations (in two variables).
Summary:
b2 ac

b
/x
= A = 0.
/y
a

To get independent of , D has to be different from D , so C =

6 0 in this case, but B is
now given by




 2

b
b


b
B = a D + b + D + c
= +c
,
a
a
y y
a
y y
so that B = 0. When b2 ac = 0 the PDE has only one family of characteristic curves, for
(x, y) = constant, whose equation is now
dy
b
= .
dx
a
Thus we have to consider three different cases.

(3.6)

(via the chain rule);

2u
2u
+
+ (lower order terms) = 0.
2 2

(3.4)

Similarly we find that the characteristic curves described by (x, y) = constant satisfy

dy
b b2 ac
=
.
(3.5)
dx
a

D =

40

Canonical form
Type

>0
2u

+ ... = 0

Hyperbolic

=0
2u
2

+ ... = 0

Parabolic

<0
2u
2

2u
2

+ ... = 0

Elliptic

E.g.
The wave equation,

2u
2u
c2w
= 0,
2
t
x2
is hyperbolic (b2 ac = c2w > 0) and the two families of characteristics are described
by dx/dt = cw i.e. = x cw t and = x + cw t. So, the equation transforms into its
canonical form 2 u/ = 0 whose solutions are waves travelling in opposite direction
at speed cw .

Chapter 3 Second Order Linear and Semilinear Equations in Two Variables

41

42

3.1 Classification and Standard Form Reduction

The diffusion (heat conduction) equation,

2 u 1 u

= 0,
x2 t
is parabolic
constant.

(b2

is elliptic

u = f () + 2 g(),

We need to impose two conditions on u or its partial derivatives to determine the functions
f and g i.e. to find a particular solution.

ac = 1 < 0).

2u 2u
y 2 + 2 = 0,
x
y

(b ac = 0 y = y)

1 2u 2u
+ 2 = 0,
1 m2 x2
y

(b2 ac =

1
)
1 m2

is elliptic if m < 1 and hyperbolic if m > 1.

y2

y3

u
u
+ x3
x
y

We can choose arbitrarily provided and are independent. We choose = y. (Exercise,

try it with = x.) Then

2u
2u
2u
= 4xy 2 + 2x
,
xy

and the equation becomes

2y 2

2u

2u

1. directly:

a=
Here b = xy
so b2 ac = (xy)2 x2 y 2 = 0 parabolic equation.

c = x2
On = constant,

dy
b + b2 ac
b
x
=
= = = x2 + y 2 .
dx
a
a
y

u
u u
= 2y
+
,
y

Reduce to canonical form and then solve

2u
u
2u
2 2 + 1 = 0 in 0 x 1, y > 0, with u =
= x on y = 0.
xy
y
y

a=1
Here b = 1/2
so b2 ac = 9/4 (> 0) equation is hyperbolic.

c = 2
Characteristics:


dy
1 3
/x
/x
.
= = 1 or 2
=
or
dx
2 2
/y
/y

u
u
= 2x ,
x

2u

Two methods of solving:

Reduce to the canonical form

2u
1
2u
2u
2xy
+ x2 2 =
2
x
xy
y
xy

Example 2:
x2

is elliptic in y > 0, parabolic for y = 0 and hyperbolic in y < 0, and for small disturbances
in incompressible (inviscid) flow

y2

where f and g are arbitrary functions (via integrating factor method), i.e.
u = f (x2 + y 2 ) + y 2 g(x2 + y 2 ).

2u 2u
+ 2 = 0,
x2
y

The type of a PDE is a local property. So, an equation can change its form in different regions
of the plane or as a parameter is changed. E.g. Tricomis equation

Example 1:

(canonical form)

This has solution

ac = 0). The characteristics are given by dt/dx = 0 i.e. = t =

Laplaces equation,
(b2

2 u 1 u

= 0.
2

i.e.

u
2u
2u
=2
+ 4x2 2 ,
x2

2u
2u
2u
u
2u
+ 4y 2 2 + 4y
+ 2,
=2
y 2

dy
1
= 2 x y = constant
dx
2

and

dy
= 1 x + y = constant.
dx

2. simultaneous equations:

y )
=2
x=
=x
x
y
2

=x+y
y=

=
x
y

So,

u
u u
=
+
,
x

u
1 u u
=
+
,
y
2

2u
1 2u 1 2u
2u
+
=
+ 2,
2
xy
2
2

1
( + 2)
3
2
( )
3

2u
u
2u
2u
=
+2
+ 2
2
2
x

2u
1 2u
2u
2u
=

+ 2,
2
2
y
4

and the equation becomes

2u

2u

u
u
+ 4x2 y 2 2 8x2 y 2 2 4x2 y
+ 2x2
+ 4x2 y 2 2



2
2
1
u
u
2 u
2 u
3 u
+ 4x y
=
+x
+ 2x3 y
+ x3
2xy
,
2

xy

2u
u
2u
2u 1 2u 1 2u
2u 1 2u
2u
+2
+
+2
+ 2
+ 2
2 2 + 1 = 0,
2

2 2
2

2 2

9 2u
+ 1 = 0,
2

canonical form.

Chapter 3 Second Order Linear and Semilinear Equations in Two Variables

43

So 2 u/ = 2/9 and general solution is given by

44

3.2 Extensions of the Theory

To obtain a real transformation, put

2
u(, ) = + f () + g(),
9

= + = 2y x

where f and g are arbitrary functions; now, we need to apply two conditions to determine
these functions.
When, y = 0, = = x so the condition u = x at y = 0 gives
2
2
u( = x, = x) = x2 + f (x) + g(x) = x f (x) + g(x) = x + x2 .
9
9

(3.7)

So,

and = i( ) = x 3.

u
u u
=
+ 3
,
x

u
u
2u
=
=2
,
y

x2
2
2
2
u
u
u
+2 3
= 2
,
xy
2

and the equation transforms to

2u
2u
2u
2 3
,
+3
2

2
2u
2u
=4
,
2
y
2

2u
2u
2u
2u
2u
2u
2 3
2
+2 3
= 0.
+3
+4
2
2
2

Also, using the relation

u
1 u u
1
1
2
=
+
= f 0 () + g 0 (),
y
2

9
2
9

2u
2u
+
= 0,
2
2

canonical form.

the condition u/y = x at y = 0 gives

2
1
10
u
1
1
( = x, = x) = x f 0 (x) x + g 0 (x) = x g 0 (x) f 0 (x) =
x,
y
9
2
9
2
9
5
1
(3.8)
and after integration, g(x) f (x) = x2 + k,
2
9
where k is a constant. Solving equation (3.7) and equation (3.8) simultaneously gives,
2
2
2
f (x) = x x2 k
3
9
3
or, in terms of and

f () =

and

g(x) =

2
2
2
2 k
3
9
3

2
1
4
x + x2 + k,
3
9
3

and g() =

2
1
4
+ 2 + k.
3
9
3

So, full solution is

1
2
2
2
4
u(, ) = + 2 + + 2 ,
9
3
9
3
9
1
2
= (2 + ) + ( )(2 + ).
3
9
y2
u(x, y) = x + xy + .
2
Example 3:

(check this solution.)

Reduce to canonical form

2u
2u
2u
+
+ 2 = 0.
x2 xy
y

a=1
so b2 ac = 3/4 (< 0) equation is elliptic.
Here b = 1/2

c=1

Find and via

)
)

dy/dx = (1 + i 3)/2
= y 12 (1 + i 3)x

= constant on dy/dx = (1 i 3)/2

= y 12 (1 i 3)x

= constant on

3.2
3.2.1

Extensions of the Theory

Linear second order equations in n variables

There are two obvious ways in which we might wish to extend the theory.
To consider quasilinear second order equations (still in two independent variables.) Such
equations can be classified in an invariant way according to rule analogous to those developed
above for linear equations. However, since a, b and c are now functions of x u, y u and u its
type turns out to depend in general on the particular solution searched and not just on the
values of the independent variables.
To consider linear second order equations in more than two independent variables. In such
cases it is not usually possible to reduce the equation to a simple canonical form. However,
for the case of an equation with constant coefficients such a reduction is possible. Although
this seems a rather restrictive class of equations, we can regard the classification obtained as
a local one, at a particular point.
Consider the linear PDE
n
X

aij

i,j=1

X u
2u
+
bi
+ cu = d.
xi xj
xi
i=1

Without loss of generality we can take the matrix A = (a ij ), i, j = 1 n, to be symmetric

(assuming derivatives commute). For any real symmetric matrix A, there is an associate
orthogonal matrix P such that P T AP = where is a diagonal matrix whose element are
the eigenvalues, i , of A and the columns of P the linearly independent eigenvectors of A,
ei = (e1i , e2i , , eni ). So
P = (eij ) and = (i ij ), i, j = 1, , n.
Now consider the transformation x = P , i.e. = P 1 x = P T x (P orthogonal) where
x = (x1 , x2 , , xn ) and = (1 , 2 , , n ); this can be written as
xi =

n
X
j=1

eij j

and

j =

n
X
i=1

eij xi .

Chapter 3 Second Order Linear and Semilinear Equations in Two Variables

So,

X u
u
=
eik
xi
k

45

n
X
u
2u
=
eik ejr .
xi xj
k r

and

k=1

k,r=1

The original equation becomes,

n
n
X
X

i,j=1 k,r=1

u
aij eik ejr + (lower order terms) = 0.
k r

But by definition of the eigenvectors of A,

eTk

n
X

A er =

i,j=1

46

3.2 Extensions of the Theory

which takes prescribed values on a given curve which we assume is represented parametrically in the form
x = (), y = (),
for I, where I is an interval, 0 1 , say. (Usually consider piecewise smooth
curves.)
We specify Cauchy data on : u, u/x and u/y are given I, but note that
we cannot specify all these quantities arbitrarily. To show this, suppose u is given on by
u = f (); then the derivative tangent to , du/d, can be calculated from du/d = f 0 ()
but also
du
u dx u dy
=
+
,
d
x d y d
u
u
+ 0 ()
= f 0 (),
= 0 ()
x
y

eik aij ejr r rk .

Then equation simplifies to

n
X
k=1

u
k 2 + (lower order terms) = 0.
k

so, on , the partial derivatives u/x, u/y and u are connected by the above relation.
Only derivatives normal to and u can be prescribed independently.

Equation is elliptic if and only if all k are non-zero and have the same sign. E.g.
Laplaces equation
2u 2u 2u
+ 2 + 2 = 0.
x2
y
z

So, the Cauchy problem consists in finding the solution u(x, y) which satisfies the following
conditions

u((), ()) = f ()

,
u

((), ()) = g()

and
n

When all the k are non-zero and have the same sign except for precisely one of them,
the equation is hyperbolic. E.g. the wave equation

Set, p = u/x and q = u/y so that on , p and q are known; then

We are now in a position to classify the equation.

2u
c2
t2

2u 2u 2u
+ 2+ 2
x2
y
z

= 0.

When all the k are non-zero and there are at least two of each sign, the equation
ultra-hyperbolic. E.g.
2u 2u
2u 2u
+ 2 =
+
;
2
x1 x2
x23 x24
such equation do not often arise in mathematical physics.
If any of the k vanish the equation is parabolic. E.g. heat equation
u

3.2.2

2u 2u 2u
+ 2 + 2
x2
y
z

= 0.

The Cauchy Problem

Consider the problem of finding the solution of the equation

2u
2u
2u
+ c 2 + F (x u, y u, u, x, y) = 0
a 2 + 2b
x
xy
y

where I and /n = n denotes a normal derivative to (e.g. n = [ 0 , 0 ]T ); the

partial derivatives u/x and u/y are uniquely determined on by these conditions.

dp
2 u dx
2 u dy
=
+
2
ds
x ds xy ds

and

dq
2 u dx 2 u dy
=
+ 2 .
ds
xy ds
y ds

Combining these two equations with the original PDE gives the following system of equations
for 2 u/x2 , 2 u/xy and 2 u/y 2 on (in matrix form),

2u
x2
2u
xy
2u
y 2

dp

ds

dq
ds

2b

dx
where M =
ds

dy
ds
dx
ds

0 .

dy
ds

So, if det(M) 6= 0 we can solve the equations uniquely and find 2 u/x2 , 2 u/xy and
2 u/y 2 on . By successive differentiations of these equations it can be shown that the
derivatives of u of all orders are uniquely determined at each point on for which det(M) 6= 0.
The values of u at neighbouring points can be obtained using Taylors theorem.
So, we conclude that the equation can be solved uniquely in the vicinity of provided
det(M) 6= 0 (Cauchy-Kowaleski theorem provides a majorant series ensuring convergence of
Taylors expansion).

Chapter 3 Second Order Linear and Semilinear Equations in Two Variables

47

Consider what happens when det(M) = 0, so that M is singular and we cannot solve uniquely
for the second order derivatives on . In this case the determinant det(M) = 0 gives,
a

dy
ds

2

2b

dx dy
+c
ds ds

dx
ds

2

= 0.

But,
dy
dy/ds
=
dx
dx/ds
and so (dividing through by dx/ds), dy/dx satisfies the equation,

 2
dy
dy
dy
b b2 ac
2b
a
+ c = 0,
i.e.
=
dx
dx
dx
a

or

dy
b
= .
dx
a

The exceptional curves , on which, if u and its normal derivative are prescribed, no unique
solution can be found satisfying these conditions, are the characteristics curves.

48

3.2 Extensions of the Theory