MA 201: Lecture - 2

Methods of characteristics

MA201(2017):PDE
Example
Let us warm up with a simple example

ux = u + c, c is function of x, y . (1)

MA201(2017):PDE
Example
Let us warm up with a simple example

ux = u + c, c is function of x, y . (1)

Observe
Equation (1) contains no derivative with respect to the y variable,
we can regard this variable as a parameter.
Thus, for fixed y , we are actually dealing with an ODE, the solution
is immediate:
 x
Z
x

u(x, y ) = e e c(, y )d + T (y ) . (2)
0

Suppose, we supplement (1) with the initial condition u(0, y ) = y .

Then the unique solution is given by
Z x
u(x, y ) = e x e c(, y )d + y .
 
(3)
0

MA201(2017):PDE
Example
Consider the following IVP :

ux = u, & u(x, 0) = 2x. (4)

MA201(2017):PDE
Example
Consider the following IVP :

ux = u, & u(x, 0) = 2x. (4)

The solution of (4) now becomes u(x, y ) = e x T (y ) and with the

condition u(x, 0) = 2x, we must have T (0) = 2xe x , which is of
course impossible.

MA201(2017):PDE
Example
Consider the following IVP :

ux = u, & u(x, 0) = 2x. (4)

The solution of (4) now becomes u(x, y ) = e x T (y ) and with the

condition u(x, 0) = 2x, we must have T (0) = 2xe x , which is of
course impossible.
Consider following IVP:

ux = u, & u(x, 0) = 2e x . (5)

MA201(2017):PDE
Example
Consider the following IVP :

ux = u, & u(x, 0) = 2x. (4)

The solution of (4) now becomes u(x, y ) = e x T (y ) and with the

condition u(x, 0) = 2x, we must have T (0) = 2xe x , which is of
course impossible.
Consider following IVP:

ux = u, & u(x, 0) = 2e x . (5)

Now T (y ) should satisfy T (0) = 2. Thus every function T (y )

satisfying T (0) = 2 will provide a solution for the equation together
with the initial condition. Hence, the IVP has infinitely many
solutions.

MA201(2017):PDE
 Thus we have seen so far an example in which a problem had a
unique solution, an example where there was no solution at all, and
an example where there are infinitely many solutions.

MA201(2017):PDE
 Thus we have seen so far an example in which a problem had a
unique solution, an example where there was no solution at all, and
an example where there are infinitely many solutions.
Well-posed Problem(In the sense of Hadamard)
A problem (PDE + side condition) is said to be well-posed if it satisfies
the following criteria:
1 The solution must exist.
2 The solution should be unique.
3 The solution should depend continuously on the initial and/or
boundary data.
If one or more of the conditions above does not hold, we say that the
problem is ill-posed.
Remark: The well-posedness of a PDE depends not only its structure
but also on the given initial/boundary conditions.

MA201(2017):PDE
Method of characteristics:

Motivation: Consider the Transport equation:

u u
+ = 0.
t x

Let C (s) = (x(s), t(s)) be a smooth curve in the xt-plane defined

parametrically with a parameter s J, where J R is a closed set
such that the direction of tangent at any point (x(s), t(s)) given by
( dx(s) dt(s)
ds , ds ) = (1, 1).
As s varies over J, the values of (x(s), t(s)) change resulting in the
change of the value of u(x(s), t(s)).

MA201(2017):PDE
 Mathematically, this change in the value of u(x(s), t(s)) is given by

d u dx u dt
u ((x(s), t(s))) = +
ds x ds t ds
= ux (x(s), t(s)) + ut (x(s), t(s))
= 0.

Thus u(x(s), t(s)) is constant along the curve C (s).

If the value of u(x(s0 ), t(s0 )), s0 J is given for some point
(x(s0 ), t(s0 )) on the curve C (s), then we get the value of
u(x(s), t(s)) for all points on C (s).
The curve C (s) is called characteristic curve.
Remark: The problem of finding the characteristic curve C (s) is
equivalent to solving a system of ODEs.

dx(s) dt(s)
= 1, = 1.
ds ds

MA201(2017):PDE
Method of characteristics for Linear First-Order PDEs
Consider the following first-order linear PDE

a(x, y )ux + b(x, y )uy = c(x, y )u + d(x, y ), (6)

where a, b, c, and d are continuously differentiable functions of x and y .
Definition (Solution of a PDE)
A function u(x, y ) is a solution to (6) if u and its partial derivatives
appearing in (6) satisfy (6) identically for (x, y ) in some region R2 .
Thus, for a solution to (6), we are looking for a function u of x and
y which satisfies equation (6). In such case solution is given
explicitly in terms of independent variables x and y .
In more general terms, we are looking for an expression
F (x, y , u) = 0 involving x, y and u which leads to equation (6). In
such case, we get an implicit form of the solution.
Observe that u = u(x, y ) means u is function of x and y or
F (x, y , u) = 0 gives a surface in R3 . This is known as integral
surface or solution surface.

MA201(2017):PDE
Observe that the left hand side of (6), i.e.,

a(x, y )ux + b(x, y )uy = u (a(x, y ), b(x, y ))

is (essentially) a directional derivative of u(x, y ) in the direction of the

vector (a(x, y ), b(x, y )), where (a(x, y ), b(x, y )) is defined and nonzero.
Remark: When a(x, y ) = a and b(x, y ) = b are constants, the vector
(a(x, y ), b(x, y )) = (a, b) had a fixed direction and magnitude, but now
the vector can change as its base point (x, y ) varies. Thus,
(a(x, y ), b(x, y )) defines a vector field on the plane.

MA201(2017):PDE
The equations
dx dy
= a(x, y ), = b(x, y ), (7)
dt dt
determine a family of curves x = x(t), y = y (t) whose tangent vector at
dy
any point (x, y ) is given by ( dx
dt , dt ) and coincides with the direction of
the vector (a(x, y ), b(x, y )).
Therefore,
d u dx u dy
u ((x(t), y (t))) = +
dt x dt y dt
= ux (x(t), y (t))a(x(t), y (t)) + uy (x(t), y (t))b(x(t), y (t))
= c(x(t), y (t))u(x(t), y (t)) + d(x(t), y (t))
= c(t)u(t) + d(t),

where we have used the chain rule and (6).

MA201(2017):PDE
Thus, along these curves, u(t) = u(x(t), y (t)) satisfies the ODE

u 0 (t) c(t)u(t) = d(t). (8)

h R i
t
Let (t) = exp 0 c( )d be an integrating factor for (8). Then, the
solution along this curve is given by
Z t 
1
u(t) = ( )d( )d + u(0) . (9)
(t) 0

MA201(2017):PDE
Remarks:
The system of ODEs (7) is known as the characteristics equation for
the PDE (6). The solution curves of the characteristic equation are
the characteristics curves for (6).
The surface u = u(x, y ) is called integral surface for (6).
The values u(t) of the solution u along the entire characteristics
curve are completely determined, once the value
u(0) = u(x(0), y (0)) is prescribed.
Assuming certain smoothness conditions on the functions a, b, c,
and d, the existence and uniqueness theory for ODEs guarantees a
unique solution curve (x(t), y (t), u(t)) of (7)-(8).
The characteristics are paths in the solution domain along which
information propagates. Thus these curves control the flow of the
information.

MA201(2017):PDE
The method of characteristics for an IVP
Let the initial curve be given parametrically as:

x = x(s), y = y (s), for s J R (10)

Every value of s fixes a point on through which a unique characteristic

curve passes. The family of characteristic curves determined by the
points of may be parameterized as

x = x(t, s), y = y (t, s),

and are given as the solutions x(t, s) and y (t, s) of the characteristics
system (for each fixed s)

d d
x(t, s) = a(x(t, s), y (t, s)), y (t, s) = b(x(t, s), y (t, s)) (11)
dt dt
with given initial values x(0, s) and y (0, s). (Note that t = 0
corresponding to the initial curve . )

MA201(2017):PDE
 u Characteristics
Curves

Figure: Characteristic curves and construction of the integral surface

MA201(2017):PDE
Suppose that
u(x(0, s), y (0, s)) = g (s), (12)
where g (s) is a given function. We obtain u(x(t, s), y (t, s)) as follows:
Let

u(t, s) = u(x(t, s), y (t, s)), c(t, s) = c(x(t, s), y (t, s)),
d(t, s) = d(x(t, s), y (t, s)).

and  Z t 
(t, s) = exp c(t, s)dt . (13)
0

MA201(2017):PDE
Analogous to formula (9), for each fixed s, we obtain
Z t 
1
u(t, s) = (t, s)d(t, s)dt + g (s) . (14)
(t, s) 0

u(t, s) is the value of u at the point (x(t, s), y (t, s)) on the characteristic
curves.
Note: As s and t vary, the point (x, y , u), in xyu-space, given by

x = x(t, s), y = y (t, s), u = u(t, s), (15)

traces out the surface of the graph of the solution u of the PDE (6)
which meets the initial curve (12).
The equations (15) constitute the parametric form of the solution of (6)
satisfying the initial condition (12) (i.e., a surface in (x, y , u)-space that
contains the initial curve )

MA201(2017):PDE
Remarks.
By implicit function theorem, if the Jacobian

(x, y )
J =
(t, s)
x y y x
=
t s t s 
 a b 
 6= 0
= 
 (x(0, s))s (y (0, s))s  (16)

on , then x = x(t, s) and y = y (t, s) can be inverted to give s and

t as (smooth) functions of x and y i.e., s = s(x, y ) and t = t(x, y ).
The resulting function U(x, y ) = u(t(x, y ), s(x, y )) satisfies the
PDE (6) in a neighborhood of the curve (in view of (8) and the
initial condition (10)) and is the unique solution of the IVP.
The condition (16) is called transversality condition.
Thus the method of characteristics gives a unique solution of the
Cauchy problem provided (16) holds.

MA201(2017):PDE
Example
Determine the solution the following IVP:
u u
+c = 0, u(x, 0) = f (x),
y x

where f (x) is a given function and c is a constant.

Solution.
Step 1.(Finding characteristic curves)
To apply the method of characteristics, parameterize the initial curve
C as follows: as follows:
x = s, y = 0, u = f (s). (17)
The family of characteristics curves x((t, s), y (t, s)) are determined
by solving the ODEs
d d
x(t, s) = c, y (t, s) = 1
dt dt
The solution of the system is
x(t, s) = ct + c1 (s) and y (t, s) = t + c2 (s).
MA201(2017):PDE
 Step 2. (Applying IC)
Using the initial conditions

x(0, s) = s, y (0, s) = 0.

we find that
c1 (s) = s, c2 (s) = 0,
and hence
x(t, s) = ct + s and y (t, s) = t.
Step 3. (Writing the parametric form of the solution)
Comparing with (6), we have c(x, y ) = 0 and d(x, y ) = 0.
Therefore, using (13) and (14), we find that

d(t, s) = 0, (t, s) = 1.

Since u(x(0, s), y (0, s)) = u(s, 0) = g (s) = f (s), we obtain

u(t, s) = f (s). Thus, the parametric form of the solution of the
problem is given by

x(t, s) = ct + s, y (t, s) = t, u(t, s) = f (s).

MA201(2017):PDE
 Step 4. (Expressing u(s, t) in terms of U(x, y ))
Expressing s and t as s = s(x, y ) and t = t(x, y ), we have

s = x cy , t = y .

We now write the solution in the explicit form as

U(x, y ) = u(t(x, y ), s(x, y )) = f (x cy ).

Clearly, if f (x) is differentiable, the solution U(x, y ) = f (x cy )

satisfies given PDE as well as the initial condition.

MA201(2017):PDE