4

The one-dimensional wave equation

4.1 Introduction
In this chapter we study the one-dimensional wave equation on the real line. The
canonical form of the wave equation will be used to show that the Cauchy problem
is well-posed. Moreover, we shall derive simple explicit formulas for the solutions.
We also discuss some important properties of the solutions of the wave equation
which are typical for more general hyperbolic problems as well.

4.2 Canonical form and general solution

The homogeneous wave equation in one (spatial) dimension has the form
u tt − c2 u x x = 0 − ∞ ≤ a < x < b ≤ ∞, t > 0, (4.1)
where c ∈ R is called the wave speed, a terminology that will be justified in the
discussion below.
To obtain the canonical form of the wave equation, define the new variables
ξ = x + ct η = x − ct,
and set w(ξ, η) = u(x(ξ, η), t(ξ, η)) (see Section 3.3 for the method to obtain
this canonical transformation). Using the chain rule for the function u(x, t) =
w(ξ (x, t), η(x, t)), we obtain
u t = wξ ξt + wη ηt = c(wξ − wη ), u x = wξ ξx + wη ηx = wξ + wη ,
and
u tt = c2 (wξ ξ − 2wξ η + wηη ), u x x = wξ ξ + 2wξ η + wηη .
Hence,
u tt − c2 u x x = −4c2 wξ η = 0.

76
 4.2 General solution 77

This is the canonical formfor the wave equation. Since (wξ )η = 0, it follows that
wξ = f (ξ ), and then w = f (ξ ) dξ + G(η). Therefore, the general solution of the
equation wξ η = 0 has the form
w(ξ, η) = F(ξ ) + G(η),
where F, G ∈ C 2 (R) are two arbitrary functions. Thus, in the original variables,
the general solution of the wave equation is
u(x, t) = F(x + ct) + G(x − ct). (4.2)
In other words, if u is a solution of the one-dimensional wave equation, then there
exist two real functions F, G ∈ C 2 such that (4.2) holds. Conversely, any two
functions F, G ∈ C 2 define a solution of the wave equation via formula (4.2).
For a fixed t0 > 0, the graph of the function G(x − ct0 ) has the same shape as
the graph of the function G(x), except that it is shifted to the right by a distance
ct0 . Therefore, the function G(x − ct) represents a wave moving to the right with
velocity c, and it is called a forward wave. The function F(x + ct) is a wave
traveling to the left with the same speed, and it is called a backward wave. Indeed
c can be called the wave speed.
Equation (4.2) demonstrates that any solution of the wave equation is the sum
of two such traveling waves. This observation will enable us to obtain graphical
representations of the solutions (the graphical method).
We would like to extend the validity of (4.2). Observe that for any two real
piecewise continuous functions F, G, (4.2) defines a piecewise continuous function
u that is a superposition of a forward wave and a backward wave traveling in opposite
directions with speed c. Moreover, it is possible to find two sequences of smooth
functions, {Fn (s)}, {G n (s)}, converging at any point to F and G, respectively, which
converge uniformly to these functions in any bounded and closed interval that does
not contain points of discontinuity. The function
u n (x, t) = Fn (x + ct) + G n (x − ct)
is a proper solution of the wave equation, but the limiting function u(x, t) = F(x +
ct) + G(x − ct) is not necessarily twice differentiable, and therefore might not be
a solution. We call a function u(x, t) that satisfies (4.2) with piecewise continuous
functions F, G a generalized solution of the wave equation.
Let us further discuss the general solution (4.2). Consider the (x, t) plane. The
following two families of lines
x − ct = constant, x + ct = constant,
are called the characteristics of the wave equation (see Section 3.3). For the wave
equation, the characteristics are straight lines in the (x, t) plane with slopes ±1/c.
78 The one-dimensional wave equation

It turns out that as for first-order PDEs, the “information” is transferred via these
curves.
We arrive now at one of the most important properties of the characteristics.
Assume that for a fixed time t0 , the solution u is a smooth function except at one
point (x0 , t0 ). Clearly, either F is not smooth at x0 + ct0 , and/or the function G is
not smooth at x0 − ct0 . There are two characteristics that pass through the point
(x0 , t0 ); these are the lines
x − ct = x0 − ct0 , x + ct = x0 + ct0 .
Consequently, for any time t1 = t0 the solution u is smooth except at one or two
points x± that satisfy
x− − ct1 = x0 − ct0 , x+ + ct1 = x0 + ct0 .
Therefore, the singularities (nonsmoothness) of solutions of the wave equation
are traveling only along characteristics. This phenomenon is typical of hyperbolic
equations in general: a singularity is not smoothed out; rather it travels at a finite
speed. This is in contrast to parabolic and elliptic equations, where, as will be shown
in the following chapters, singularities are immediately smoothed out.
Example 4.1 Let u(x, t) be a solution of the wave equation
u tt − c2 u x x = 0,
which is defined in the whole plane. Assume that u is constant on the line x = 2 + ct.
Prove that u t + cu x = 0.

The solution u(x, t) has the form u(x, t) = F(x + ct) + G(x − ct). Since u(2 +
ct, t) = constant, it follows that
F(2 + 2ct) + G(2) = constant.
Set s = 2 + 2ct, we have F(s) = constant. Consequently u(x, t) = G(x − ct).
Computing now the expression u t + cu x , we obtain
u t + cu x = −cG (x − ct) + cG (x − ct) = 0.

4.3 The Cauchy problem and d’Alembert’s formula

The Cauchy problem for the one-dimensional homogeneous wave equation is given
by
u tt − c2 u x x = 0 − ∞ < x < ∞, t > 0, (4.3)
u(x, 0) = f (x), u t (x, 0) = g(x), −∞ < x < ∞. (4.4)
 4.3 The Cauchy problem and d’Alembert’s formula 79

A solution of this problem can be interpreted as the amplitude of a sound wave

propagating in a very long and narrow pipe, which in practice can be considered as
a one-dimensional infinite medium. This system also represents the vibration of an
infinite (ideal) string. The initial conditions f, g are given functions that represent
the amplitude u, and the velocity u t of the string at time t = 0.
A classical (proper) solution of the Cauchy problem (4.3)–(4.4) is a function
u that is continuously twice differentiable for all t > 0, such that u and u t are
continuous in the half-space t ≥ 0, and such that (4.3)–(4.4) are satisfied. Generally
speaking, classical solutions should have the minimal smoothness properties in
order to satisfy continuously all the given conditions in the classical sense.
Recall that the general solution of the wave equation is of the form

u(x, t) = F(x + ct) + G(x − ct). (4.5)

Our aim is to find F and G such that the initial conditions of (4.4) are satisfied.
Substituting t = 0 into (4.5) we obtain

u(x, 0) = F(x) + G(x) = f (x). (4.6)

Differentiating (4.5) with respect to t and substituting t = 0, we have

u t (x, 0) = cF (x) − cG (x) = g(x). (4.7)

Integration of (4.7) over the integral [0, x] yields


1 x
F(x) − G(x) = g(s) ds + C, (4.8)
c 0
where C = F(0) − G(0). Equations (4.6) and (4.8) are two linear algebraic equa-
tions for F(x) and G(x). The solution of this system of equations is given by
 x
1 1 C
F(x) = f (x) + g(s) ds + , (4.9)
2 2c 0 2
 x
1 1 C
G(x) = f (x) − g(s) ds − . (4.10)
2 2c 0 2
By substituting these expressions for F and G into the general solution (4.5),
we obtain the formula
 x+ct
f (x + ct) + f (x − ct) 1
u(x, t) = + g(s) ds, (4.11)
2 2c x−ct
which is called d’Alembert’s formula. Note that sometimes (4.9)–(4.10) are also
useful, as they give us explicit formulas for the forward and the backward waves.
The following examples illustrate the use of d’Alembert’s formula.
80 The one-dimensional wave equation

Example 4.2 Consider the Cauchy problem

u tt − u x x = 0 − ∞ < x < ∞, t > 0,


 0 −∞ < x < −1,

x + 1 −1 ≤ x ≤ 0,
u(x, 0) = f (x) =

 1 − x 0 ≤ x ≤ 1,

0 1 < x < ∞,

0 −∞ < x < −1,
u t (x, 0) = g(x) = 1 −1 ≤ x ≤ 1,

0 1 < x < ∞.

(a) Evaluate u at the point (1, 12 ).

(b) Discuss the smoothness of the solution u.

(a) Using d’Alembert’s formula, we find that

f ( 32 )+ f ( 12 )  3
u(1, 12 ) = 2
+ 1
2 1
2
g(s) ds.
2

Since 3
2
> 1 it follows that f ( 32 ) = 0. On the other hand, 0 ≤ 12 ≤ 1; therefore,
3 1
f ( 12 ) = 12 . Evidently, 12 g(s)ds = 1 1ds = 12 . Thus, u(1, 12 ) = 12 .
2 2

(b) The solution is not classical, since u ∈ C 1 . Yet u is a generalized solution of

the problem. Note that although g is not continuous, nevertheless the solution u
is a continuous function. The singularities of the solution propagate along charac-
teristics that intersect the initial line t = 0 at the singularities of the initial condi-
tions. These are exactly the characteristics x ± t = −1, 0, 1. Therefore, the solu-
tion is smooth in a neighborhood of the point (1, 12 ) which does not intersect these
characteristics.
Example 4.3 Let u(x, t) be the solution of the Cauchy problem
u tt − 9u x x = 0 − ∞ < x < ∞, t > 0,

1 |x| ≤ 2,
u(x, 0) = f (x) =
0 |x| > 2,

1 |x| ≤ 2,
u t (x, 0) = g(x) =
0 |x| > 2.

(a) Find u(0, 16 ).

(b) Discuss the large time behavior of the solution.
 4.3 The Cauchy problem and d’Alembert’s formula 81

(c) Find the maximal value of u(x, t), and the points where this maximum is achieved.
(d) Find all the points where u ∈ C 2 .

(a) Since

f (x + 3t) + f (x − 3t) 1 x+3t
u(x, t) = + g(s)ds,
2 6 x−3t

it follows that for x = 0 and t = 16 , we have

 1  1
1 f ( 12 ) + f (− 12 ) 1 2 1+1 1 2 7
u(0, ) = + g(s) ds = + 1ds = .
6 2 6 − 12 2 6 − 12 6
(b) Fix ξ ∈ R and compute limt→∞ u(ξ, t). Clearly,
 ξ +3t  2
lim f (ξ + 3t) = 0, lim f (ξ − 3t) = 0, lim g(s) ds = 1ds = 4.
t→∞ t→∞ t→∞ ξ −3t −2

Therefore, limt→∞ u(ξ, t) = 23 .

(c) Recall that for any real functions f, g,
max{ f (x) + g(x)} ≤ max f (x) + max g(x).
It turns out that in our special case there exists a point (x, t), where all the terms in (4.11)
attain their maximal value simultaneously, and therefore at such a point the maximum
of u is attained.
Indeed, max{ f (x + 3t)} = 1 which is attained on the strip −2 ≤ x + 3t ≤ 2. Sim-
ilarly, max{ f (x − 3t)} 2= 1 which is attained on the strip −2 ≤ x − 3t ≤ 2, while
x+3t
max{ x−3t g(s) ds = −2 1ds = 4, and it is attained on the intersection of the half-
planes x + 3t ≥ 2 and x − 3t ≤ −2. The intersection of all these sets is the set of all
points that satisfy the two equations
x + 3t = 2,
x − 3t = −2.

This system has a unique solution at (x, t) = (0, 23 ). Thus, the solution u achieves its
maximum at the point (0, 23 ), where u(0, 23 ) = 53 .
(d) The initial conditions are smooth except at the points x = ±2. Therefore, the solution
is smooth at all points that are not on the characteristics
x ± 3t = −2, x ± 3t = 2.
The function u is a generalized solution that is piecewise continuous for any fixed time
t > 0.

The well-posedness of the Cauchy problem follows from the d’Alembert

formula.

Theorem 4.4 Fix T > 0. The Cauchy problem (4.3)–(4.4) in the domain −∞ <
x < ∞, 0 ≤ t ≤ T is well-posed for f ∈ C 2 (R), g ∈ C 1 (R).
82 The one-dimensional wave equation

Proof The existence and uniqueness follow directly from the d’Alembert formula.
Indeed, this formula provides us with a solution, and we have shown that any
solution of the Cauchy problem is necessarily equal to the d’Alembert solution.
Note that from our smoothness assumption ( f ∈ C 2 (R), g ∈ C 1 (R)), it follows that
u ∈ C 2 (R × (0, ∞)) ∩ C 1 (R × [0, ∞)), and therefore, the d’Alembert solution is a
classical solution. On the other hand, for f ∈ C(R) and g that is locally integrable,
the d’Alembert solution is a generalized solution.
It remains to prove the stability of the Cauchy problem, i.e. we need to show that
small changes in the initial conditions give rise to a small change in the solution.
Let u i be two solutions of the Cauchy problem with initial conditions f i , gi , where
i = 1, 2. Now, if
| f 1 (x) − f 2 (x)| < δ, |g1 (x) − g2 (x)| < δ,
for all x ∈ R, then for all x ∈ R and 0 ≤ t ≤ T we have
| f 1 (x + ct) − f 2 (x + ct)| | f 1 (x − ct) − f 2 (x − ct)|
|u 1 (x, t) − u 2 (x, t)| ≤ +
2 2
 x+ct
1 1 1
+ |g1 (s) − g2 (s)| ds < (δ + δ) + 2ctδ ≤ (1 + T )δ.
2c x−ct 2 2c
Therefore, for a given ε > 0, we take δ < ε/(1 + T ). Then for all x ∈ R and 0 ≤
t ≤ T we have
|u 1 (x, t) − u 2 (x, t)| < ε.

Remark 4.5 (1) The Cauchy problem is ill-posed on the domain −∞ < x <
∞, t ≥ 0.
(2) The d’Alembert formula is also valid for −∞ < x < ∞, T < t ≤ 0, and the
Cauchy problem is also well-posed in this domain. The physical interpretation is
that the process is reversible.