Ch 10.

1:
Two-Point Boundary Value Problems
In many important physical problems there are two or more
independent variables, so the corresponding mathematical
models involve partial differential equations.
This chapter treats one important method for solving partial
differential equations, known as separation of variables.
Its essential feature is the replacement of a partial differential
equation by a set of ordinary differential equations, which
must be solved subject to given initial or boundary conditions.
Section 10.1 deals with some basic properties of boundary
value problems for ordinary differential equations.
The solution of the partial differential equation is then a sum,
usually an infinite series, formed from the solutions to the
ordinary differential equations, as we see later in the chapter.
Boundary Conditions
Up to this point we have dealt with initial value problems,
consisting of a differential equation together with suitable
initial conditions at a given point.
A typical example, as discussed in Chapter 3, is
y p(t ) y q(t ) y g (t ), y(t0 ) y0 , y(t0 ) y0
Physical applications often require the dependent variable y or
its derivative y' to be specified at two different points.
Such conditions are called boundary conditions.
The differential equation and suitable boundary conditions
form a two-point boundary value problem.
A typical example is
y p( x) y q( x) y g ( x), y( ) y0 , y( ) y1
Homogeneous Boundary Value Problems
The natural occurrence of boundary value problems usually
involves a space coordinate as the independent variable, so we
use x instead of t in the boundary value problem
y p( x) y q( x) y g ( x), y( ) y0 , y( ) y1
Boundary value problems for nonlinear equations can be
posed, but we restrict ourselves to linear equations only.
If the above boundary value problem has the form
y p( x) y q( x) y 0, y ( ) 0, y ( ) 0
then it is said to be homogeneous. Otherwise, the problem is
nonhomogeneous.
Solutions to Boundary Value Problems
To solve the boundary value problem,
y p( x) y q( x) y g ( x), y( ) y0 , y( ) y1
we need to find a function y = (x) that satisfies the differential
equation on the interval < x < and that takes on the
specified values y0 and y1 at the endpoints.
Initial value and boundary value problems may superficially
appear similar, but their solutions differ in important ways.
Under mild conditions on the coefficients, an initial value
problem is certain to have a unique solution.
Yet for similar conditions, boundary value problems may have
a unique solution, no solution, or infinitely many solutions.
In this respect, linear boundary value problems resemble
systems of linear algebraic equations.
Linear Systems
Consider the system Ax = b, where A is an n x n matrix, b is a
given n x 1 vector, and x is an n x 1 vector to be determined.
Recall the following facts (see Section 7.3):
If A is nonsingular, then Ax = b has unique solution for any b.
If A is singular, then Ax = b has no solution unless b satisfies a certain
additional condition, in which case there are infinitely many solutions.
The homogeneous system Ax = 0 always has the solution x = 0.
If A is nonsingular, then this is the only solution, but if A is singular,
then there are infinitely many (nonzero) solutions.
Thus the nonhomogeneous system has a unique solution iff the
homogeneous system has only the solution x = 0, and
the nonhomogeneous system has either no solution or infinitely
many solutions iff homogeneous system has nonzero solutions.
Example 1
Consider the boundary value problem
y 2 y 0, y (0) 1, y ( ) 0
The general solution of the differential equation is
y c1 cos 2 x c2 sin 2 x
The first boundary condition requires that c1 = 1.
From the second boundary condition, we have
c1 cos 2 c2 sin 2 0 c2 cot 2 0.2762
Thus the solution to the boundary value problem is
y cos 2 x cot 2 sin 2 x
This is an example of a nonhomogeneous boundary value
problem with a unique solution.
Example 2
Consider the boundary value problem
y y 0, y (0) 1, y ( ) a, a 0 arbitrary.
The general solution of the differential equation is
y c1 cos x c2 sin x
The first boundary condition requires that c1 = 1, while the
second requires c1 = - a. Thus there is no solution.
However, if a = -1, then there are infinitely many solutions:
y cos x c2 sin x, c2 arbitrary
This example illustrates that a nonhomogeneous boundary
value problem may have no solution, and also that under
special circumstances it may have infinitely many solutions.
Nonhomogeneous Boundary Value Problem
and Corresponding Homogeneous Problem
Corresponding to a nonhomogeneous boundary value problem
y p( x) y q( x) y g ( x), y( ) y0 , y( ) y1
is the homogeneous problem
y p( x) y q( x) y 0, y ( ) 0, y ( ) 0
Observe that this problem has the solution y = 0 for all x,
regardless of the coefficients p(x) and q(x).
This solution is often called the trivial solution and is rarely of
interest.
What we would like to know is whether the problem has other,
nonzero solutions.
Example 3
Consider the boundary value problem
y 2 y 0, y (0) 0, y ( ) 0
As in Example 1, the general solution is
y c1 cos 2 x c2 sin 2 x
The first boundary condition requires that c1 = 0.
From the second boundary condition, we have c2 = 0.
Thus the only solution to the boundary value problem is y = 0.
This example illustrates that a homogeneous boundary value
problem may have only the trivial solution y = 0.
Example 4
Consider the boundary value problem
y y 0, y (0) 0, y ( ) 0
As in Example 2, the general solution is
y c1 cos x c2 sin x
The first boundary condition requires c1 = 0, while the second
boundary condition is satisfied regardless of the value of c2.
Thus there are infinitely many solutions of the form
y c2 sin x, c2 arbitrary
This example illustrates that a homogeneous boundary value
problem may have infinitely many (nontrivial) solutions.
Linear Boundary Value Problems
Thus examples 1 through 4 illustrate(but do not prove) that
there is the same relationship between homogeneous and
nonhomogeneous linear boundary value problems as there is
between homogeneous and nonhomogeneous linear algebraic
systems.
A nonhomogeneous boundary value problem (Example 1) has
a unique solution, and the corresponding homogeneous
problem (Example 3) has only the trivial solution.
Further, a nonhomogeneous problem (Example 2) has either
no solution or infinitely many solutions, and the corresponding
homogeneous problem (Example 4) has nontrivial solutions.
Eigenvalue Problems (1 of 8)

Recall from Section 7.3 the eigenvalue problem Ax = x.

Note that x = 0 is a solution for all , but for certain , called
eigenvalues, there are nonzero solutions, called eigenvectors.
The situation is similar for boundary value problems.
Consider the boundary value problem
y y 0, y (0) 0, y ( ) 0
This is the same problem as in Example 3 if = 2, and is the
same problem as in Example 4 if = 1.
Thus the above boundary value problem has only the trivial
solution for = 2, and has other, nontrivial solutions for = 1.
Eigenvalues and Eigenfunctions (2 of 8)

Thus our boundary value problem

y y 0, y (0) 0, y ( ) 0
has only the trivial solution for = 2, and has other, nontrivial
solutions for = 1.
By extension of the terminology for linear algebraic systems,
the values of for which nontrivial solutions occur are called
eigenvalues, and the nontrivial solutions are eigenfunctions.
Thus = 1 is an eigenvalue of the boundary value problem
and = 2 is not.
Further, any nonzero multiple of sinx is an eigenfunction
corresponding to the eigenvalue = 1.
Boundary Value Problem for > 0 (3 of 8)

We now seek other eigenvalues and eigenfunctions of

y y 0, y (0) 0, y ( ) 0
We consider separately the cases < 0, = 0 and > 0.
Suppose first that > 0. To avoid the frequent appearance of
radical signs, let = 2, where > 0.
Our boundary value problem is then
y 2 y 0, y(0) 0, y( ) 0
The general solution is
y c1 cos x c2 sin x
The first boundary condition requires c1 = 0, while the second
is satisfied regardless of c2 as long as = n, n = 1, 2, 3, .
Eigenvalues, Eigenfunctions for > 0 (4 of 8)

We have = 2 and = n. Thus the eigenvalues of

y y 0, y (0) 0, y ( ) 0
are
1 1, 2 4, 3 9, , n n 2 ,
with corresponding eigenfunctions
y1 a1 sin x, y2 a2 sin 2 x, y3 a3 sin 3x, , yn an sin nx,
where a1, a2,, an, are arbitrary constants. Choosing each
constant to be 1, we have
y1 sin x, y2 sin 2 x, y3 sin 3x, , yn sin nx,
Boundary Value Problem for < 0 (5 of 8)

Now suppose < 0, and let = -2, where > 0.

Then our boundary value problem becomes
y 2 y 0, y(0) 0, y( ) 0
The general solution is
y c1 cosh x c2 sinh x
We have chosen cosh x and sinh x instead of e x and e- x for
convenience in applying the boundary conditions.
The first boundary condition requires that c1 = 0, and from the
second boundary condition, we have c2 = 0.
Thus the only solution is y = 0, and hence there are no negative
eigenvalues for this problem.
Boundary Value Problem for = 0 (6 of 8)

Now suppose = 0. Then our problem becomes

y 0, y (0) 0, y ( ) 0
The general solution is
y c1 x c2
The first boundary condition requires that c2 = 0, and from the
second boundary condition, we have c1 = 0.
Thus the only solution is y = 0, and = 0 is not an eigenvalue
for this problem.
Real Eigenvalues (7 of 8)

Thus the only real eigenvalues of

y y 0, y (0) 0, y ( ) 0
are n = n2 with corresponding eigenfunctions proportional to
y1 sin x, y2 sin 2 x, y3 sin 3x, , yn sin nx,
There is a possibility of complex eigenvalues, but for this
particular boundary value problem it can be shown that there
are no complex eigenvalues.
Later, in Section 11.2, we discuss an important class of
boundary problems that includes the one above.
One of the useful properties of this class of problems is that all
their eigenvalues are real.
Boundary Value Problem on [0, L] (8 of 8)

In later sections of this chapter we will often encounter this

boundary value problem on a more general interval [0, L]:
y y 0, y (0) 0, y ( L) 0
If we let = 2, > 0, as before, the general solution is
y c1 cos x c2 sin x
The first boundary condition requires c1 = 0, and the second
requires = n /L, regardless of the value of c2.
Thus, as before, the eigenvalues and eigenvectors are
n = n22/L2, yn(x) = sin(n x/L),
where the eigenfunctions yn(x) are determined only up to an
arbitrary multiplicative constant.