Lecture 5

Weston Barger

July 12, 2016

1 d’Alembert’s Solution

1.1 IVP
We will now solve our first initial value problem. Consider the problem



 x ∈ (–∞, ∞), t ∈ [0, ∞)

utt (x , t ) = c 2u xx (x , t ) u(x , 0) = f (x )


ut (x , 0) = g(x )



Here, f represents the initial position of the string and g represents it’s initial velocity.
We already know that

u(x , t ) = F (x – ct ) + G(x + ct ).

We want to find what F and G are in terms of f and g. We know that

f (x ) = u(x , 0) = F (x ) + G(x )

and

g(x ) = ut (x , 0) = –cF 0 (x ) + cG 0 (x ).

Thus, we need to solve


 f (x ) = F (x ) + G(x )
 g(x ) = –cF 0 (x ) + cG 0 (x )

1
From the second equation
1Zx
–F (x ) + G(x ) = g(y) dy + α
c 0
Our system is now

 F (x ) + G(x ) = f (x )
–F (x ) + G(x ) = c1 0x g(y) dy + α
 R

Adding and subtracting these two equation together gives

1 1 Zx 1
F (x ) = f (x ) – g(y) dy – α
2 2c 0 2
1 1 Zx 1
G(x ) = f (x ) + g(y) dy + α
2 2c 0 2
Thus,

u(x , t ) = F (x – ct ) + G(x + ct )
1 1 Z x –ct 1 1 1 Z x +ct 1
= f (x – ct ) – g(y) dy – α + f (x + ct ) + g(y) dy + α
2 2c 0 2 2 2c 0 2
1 1 Z x +ct
= (f (x – ct ) + f (x + ct )) + g(y) dy.
2 2c x –ct
This solution is known as d’Alembert’s solution to the IVP.

Example 1.1. Let us solve the IVP

2

 u(x , 0) = e –x
utt = c 2 uxx  ut (x , 0) = 0

By d’Alembert’s solution, we have that

1 –(x –ct )2
 
2
u(x , t ) = e + e –(x +ct ) .
2
The initial condition splits into two parts: one going to the left and one going to the right
at speed c.

2 Introduction to Characteristics
Chapter 10 in the textbook

2
  Characteristic curves help us to visualize how information is passed through (x , t )
from initial data u(x , 0).

 We’ll start with an example before generalizing.

Let’s consider the advection equation:

ut + cux = 0 (2.1)

Recall that f (x – ct ) is a solution for any differentiable f . This constitutes all solutions.
To see this, let us make the transformation

ξ = x – ct τ = t.

We write u(x , t ) = u(ξ, τ ). Now,

∂u ∂u ∂ξ ∂u ∂τ
= · + ·
∂x ∂ξ ∂x ∂τ ∂x
= uξ · 1 + uτ · 0
= uξ

and
∂u ∂u ∂ξ ∂u ∂τ
= · + ·
∂t ∂ξ ∂t ∂τ ∂t
= uξ · (–c) + uτ · 1
= uτ – cuξ

Plugging these expressions into (2.1) gives

uτ – cuξ + cuξ = uτ = 0.

Integrating gives
Z
uτ (ξ, τ ) dτ = f (ξ).

Thus, every solution of (2.1) is of the form

u(x , t ) = f (x – ct ).

3
Note: This is why we call it the advection equation. Every solution is a translation of the
initial wave profile u(x , 0).

Choose an initial wave profile f (x ), and choose a point x0 on the x -axis. The set

L(x0 ) = {(x , t ) | x – ct = x0 }

defines a line. Namely, the line x = x0 + ct . These lines are shown in Figure 1.

Figure 1: The characteristic lines for the advection equation.

We have that along the line L(x0 ), u has the constant value

u(x , t ) = f (x – ct ) = f (x0 ).

In other words, if u(x , t ) satisfies the advection equation, then the information f (x0 ) is
transported along the x -axis and we can track the location of f (x0 ) with x (t ) = x0 + ct .

Note: The expression for the characteristic lines x (t ) = x0 + ct i.e. x – ct = x0 comes

from the equation. So, a different PDE will have different characteristic lines.

2.1 Wave Equation Characteristics

Consider the wave equation

utt = c 2 uxx .

4
Recall that d’Alembert’s solution is
1 1 Z x +ct
u(x , t ) = (u(x – ct , 0) + u(x + ct , 0)) + u (y, 0) dy.
2 2c x –ct t
Choose a point (x , t ). We see that d’Alembert’s solution depends only on u(x , 0) and
ut (x , 0) for the interval

[x – ct , x + ct ].

Figure 2: The domain of dependence for a point (x0 , t0 ).

The gray shaded cone in Figure 2 illustrates this idea. We call this cone the domain
of dependence. The value of the solution to the wave equation at (x , t ), u(x , t ), depends
on all values of u over this cone and none outside of this cone.
Assumption
Let’s consider the special case ut (x , 0). Then our solution is simply
1
u(x , t ) = (u(x – ct , 0) + u(x + ct , 0)) .
2
This shows that u(x , t ) depends only on the initial conditions on the boundary of the
cone in Figure 2. These two boundary lines are called the characteristics of the wave
equation. Thus, the characteristics determine the domain of dependence.
Let’s turn this idea around.

5
  For a given (x0 , t0 ), what is the set of points (x , t ) such that (x0 , t0 ) is in their domain
of dependence?

This domain is known as the domain of influence. The domain of influence for a point
(x0 , t0 ) is shown as the gray section of Figure 3.

Figure 3: The domain of influence for a point (x0 , t0 ).

Let’s consider the domain of influence of an interval I = [a, b]. In this case, the domain
of influence is completely determined by the boundary points a and b, as is seen in Figure
4.

Figure 4: Domain of influence for an interval I .

6
 Characteristics may sometimes be used to analyze the behavior of a solution to the
wave equation.

Example 2.1. Let us consider the problem

 

  1 x ∈ [0, 1]

 u(x , 0) =
utt = 4uxx  0 x 6∈ [0, 1] .


ut (x , 0) = 0



√
Here, the speed of propagation is c = 4 = 2. We have the solution illustrated in Figure
5.
From this plot, we immediately read off the values of u(x , t ) for any value of (x , t ).
We can plot the solution at any t , as is done in Figure 6.

Figure 5: The different solution regions for the example problem The values in blues are
values of u(x , t ) in that region. The characteristics have slope 1/2.

7
 Figure 6: This shows u(x , t ) for varying times.

3 The Wave equation on a semi-infinite domain

3.1 Fixed End

Let us consider a string of semi-infinte extent, fixed at the left end as drawn in Figure 7.

8
 Figure 7: This figure shows a string with whose left end is fixed

The condition u(0, t ) = 0 for all t ≥ 0 is our first example of a boundary condition.
Thus, our problem is



 x ∈ (0, ∞), t ∈ [0, ∞)


u(0, t ) = 0 for all t ∈ [0, ∞)


utt = c 2 uxx , (3.1)



 u(x , 0) = f (x ) for all x ∈ (0, ∞)

ut (x , 0) = g(x ) for all x ∈ (0, ∞)



where f (x ) and g(x ) are given functions.

We shall use d’Alembert’s solution and characteristics. Recall
1 1 Z x +ct
u(x , t ) = (f (x – ct ) + f (x + ct )) + g(y) dy.
2 2c x –ct
This solution is only valid if x – ct ≥ 0. Otherwise, we’d be asking to evaluated f at
negative arguments, but f was only specified for positive values. This region is shown in
gray in Figure 8. Let us call this region R.

9
 Figure 8: This figure shows the region x < ct

Let us call the solution in the shaded region uR (x , t ). In R, we still have that

uR (x , t ) = F1 (x – ct ) + G(x + ct ).

for some functions F1 (z ) and G(z ). Using the boundary condition, we see that

uR (0, t ) = 0 = F1 (–ct ) + G(ct ) ⇒ F1 (–ct ) = –G(ct ) ⇒ F (z ) = –G(–z ).

Thus, in R, we have

uR (x , t ) = –G(–x + ct ) + G(x + ct ). (3.2)

Note: Notice that both of the arguments of G in (3.2) are valid, because both –x + ct
and x + ct are greater than 0. This is because we are considering the region R
where x < ct i.e. 0 < ct – x .

As we saw in lecture 4,
1 1 Zz
G(z ) = f (z ) + g(y) dy.
2 2c 0

10
Therefore, in the region R
1 1 Z x +ct 1 1 Z –x +ct
uR (x , t ) = f (x + ct ) + g(y) ds – f (–x + ct ) – g(y) dy
2 2c 0 2 2c 0
1 1 Z ct +x
(f (x + ct ) – f (ct – x )) + g(y) dy.
2 2c ct –x
Thus, our full solution is

 1 1 R ct +x g(y) dy
(f (x + ct ) – f (ct – x )) + 2c x < tc
 2

 ct –x
u(x , t ) =  .
 1 1 R x +ct

(f (x + ct ) + f (x – ct )) + g(y) dy x ≥ tc

2 2c x –ct

How do we see what is happening? Let us consider the odd extension of u


 u(x , t ) x ≥0
û(x , t ) = (3.3)
 –u(–x , t ) x < 0.

Note that this function satisfies the problem (3.1). Since u is the super position of two
waves traveling in opposite directions, (3.3), illustrates the effect of flipping the value of
u at the boundary.

11