Math1120 - Mathematics II

Differential Equations (IV): Homogeneous Second Order Linear DEs

So far we have learnt how to find exact solutions to separable first order DEs and linear
first order DEs. For other types of first order DEs finding exact solutions can become quite
involved (or even impossible) and so instead we will turn our attention to second order DEs.
Recall that a second order DE is one that involves the second derivative of the unknown
function. As with first order DEs we can only find analytic methods for solving second order
DEs for certain classes of such DEs.

The class of second order DEs that we are going to consider in this lecture is called the
class of “homogeneous, second order, linear DEs with constant coefficients”. Let’s try to put
this into some sort of context. Recall from DEs (III) that a first order linear DE is a DE of the
form:
P( x) y  Q( x) y  f ( x) .

As we have seen, this equation is called linear because it is the sum of terms that either don’t
involve y or only involve y or its derivatives raised to the power of one1. Extending this idea
to second order DEs gives the following definition.

A differential equation of the form

P( x) y  Q( x) y  R( x) y  f ( x)

where P( x)  0 is called a second order linear DE.

In the case where f ( x)  0 the DE is called homogeneous. Thus:

A differential equation of the form

P( x) y  Q( x) y  R( x) y  0

where P( x)  0 is called a homogeneous second order linear DE.

Further, if each of P( x) , Q( x) and R( x) are constant functions then the DE is said to have
constant coefficients. Thus:

A differential equation of the form

ay  by  cy  0

where a, b, c  and a  0 is called a homogeneous second order linear DE with

constant coefficients

1
This is a very rudimentary statement of the concept of linearity.

Page 35
 Differential Equations (IV) - Lecture Notes

§1 Homogeneous Second Order Linear DEs With Constant Coefficients

Example 1: Guess solutions to the following homogeneous second order linear DEs
with constant coefficients.
(a) y  0
(b) y  y  0

(a) What function has its second derivative zero (i.e. has zero curvature)? Clearly any
linear function satisfies this property and so we guess that the general solution to
this DE is
y( x)  Ax  B , A, B  .

(b) This DE can be rewritten as

y  y .
The solution, then, is any function that has its second derivative equal to the
function itself. This is a property of any of the exponential functions
y( x)  Ae x .
It is also a property of the exponential functions
y( x)  Be x .
Thus every function of the form
y( x)  Ae x  Be x , A, B 
will be a solution to the DE.
☼

Before looking at the method for solving a homogeneous second order linear DE with
constant coefficients, i.e. a DE of the form

ay  by  cy  0 , a  0, b, c  , (1)

there are several things that we can predict about the solution, (based on Example 1 and
previous lectures):
• The general solution will be a family of functions involving two parameters.
• The exponential function is a likely candidate for a solution.

Notice also that if both y1 ( x) and y2 ( x) are solutions to (1) then so will be
y( x)  Ay1 ( x)  By2 ( x) where A and B are arbitrary constants. To see this, let y1 ( x) and
y2 ( x) be solutions to (1), i.e.
ay1  by1  cy1  0
ay2  by2  cy2  0 .
Now let
y( x)  Ay1 ( x)  By2 ( x) .
Then
y( x)  Ay1( x)  By2 ( x)
y( x)  Ay1( x)  By2( x) .

Substituting into (1) gives:

Page 36
 Math1120 - Mathematics II

lhs  ay  by  cy
 a( Ay1  By2)  b( Ay1  By2 )  c( Ay1  By2 )
 A(ay1  by1  cy1 )  B(ay2  by2  cy2 )
0 .

These observations, along with the following theorem, enable us to develop a method
of solution for DEs of the form (1).

Theorem
For the homogeneous second order linear DE
P( x) y  Q( x) y  R( x)  0
where P( x)  0 , given any pair of solutions y1 ( x) and y2 ( x) that are not constant
multiples of each other a general solution to the DE is
y( x)  Ay1 ( x)  By2 ( x) .

Note: The proof of this theorem is beyond the scope of this course.

Example 2: Given that y  e x , y  e x , y  sinh( x) and y  cosh( x) are all solutions

to the DE
y  y  0
write down a general solution for to this DE.

From the above theorem, since none of these solutions are constant multiples of
each other, any pair of them can be combined to form a general solution to the DE.
Thus possible general solutions could be
y1 ( x)  Ae x  Be x ,
y2 ( x)  A sinh( x)  B cosh( x)
y3 ( x)  Ae x  B cosh( x) ,
and so on.

Recall that a general solution to the DE contains every solution to the DE. For
example consider the particular solution
y( x)  2sinh( x)  3cosh( x) .
This solution belongs to the general solution y2 ( x) with A  2 and B  3 . It also
belongs to the general solution y1 ( x) with A  1/ 2 and B   5 2 and to the general
solution y3 ( x) with A  2 and B  5 .
☼

Thus, initially at least, our method for trying to find the solutions to Equation (1) is to
try to find two different exponential functions that satisfy the DE and then combine them to
obtain a general solution. To this end, let

y( x)  erx (2)

Page 37
 Differential Equations (IV) - Lecture Notes

where r is some, as yet unspecified, constant. Since y  rerx and y  r 2erx for this to be a
solution of (1) we have
a  r 2erx   b  rerx   c  erx   0
(ar 2  br  c)erx  0 .
Since erx  0 for any value of r or x we conclude that (2) is a solution of (1) when

ar 2  br  c  0 . (3)

Equation (3) is called the characteristic equation for a DE of the form of equation (1).

If the characteristic equation (3) has two real solutions (i.e. b2  4ac  0 ) then we are
done since we will have found two solutions to (1) and, via the given theorem, we can
combine these solutions to produce a general solution to (1). So if the solutions to the
characteristic equation are r1 and r2 say, then the general solution to (1) is
y( x)  Aer1x  Ber2 x .

Example 3: Find the general solution to y  3 y  2 y  0 .

The characteristic equation for this homogeneous second order linear DE with
constant coefficients is
r 2  3r  2  0
which has solutions
r  1, 2 .
Thus the general solution to the DE is
y( x)  Ae x  Be2 x .
☼

If the characteristic equation (3) has only one real solution (i.e. b2  4ac  0 ) then we
will have found only one exponential solution to (1), i.e.
b
y( x)  erx where r   . (4)
2a

To find a general solution we have to find another solution (that is not a constant multiple
(4)). It turns out that another solution to (1) can be found by trying a function of the form

b
y( x)  f ( x)erx , r   . (5)
2a
Differentiating y and substituting into (1) gives

aerx  r 2 f ( x)  2rf ( x)  f ( x)   berx  rf ( x)  f ( x)   cf ( x)erx  0

f ( x)erx (ar 2  br  c)  f ( x)erx (2ar  b)  f ( x)e rx  0 .

Since ar 2  br  c  0 , 2ar  b  0 and erx  0 (5) is a solution when

f ( x)  0 . (6)

Page 38
 Math1120 - Mathematics II

Clearly f ( x)  1 satisfies (6) but this just gives us solution (4). f ( x)  x also satisfies (6)
and gives us a solution that is not a constant multiple of (4). Thus a general solution to (1)
will be
b
y( x)   Ax  B  erx , r   .
2a

Example 4: Find the general solution to y  4 y  4 y  0 .

The characteristic equation for this DE is

r 2  4r  4  0
which has only the one real solution
r  2.
Thus the general solution to the DE is
y( x)  Ae2 x  Bxe2 x .
☼

The final case to consider is that the characteristic equation (3) has complex roots, say
r    i .
Even though we are looking for real solutions, for the moment let us consider the “complex
solutions”
y1 ( x)  e  and y2 ( x)  e  .
 i x  i  x

Using Euler’s formula

ei  cos    i sin  
we can write these complex solutions as
y1 ( x)  e x ei x  e x  cos(  x)  i sin(  x) 
y2 ( x)  e xei x  e x  cos( x)  i sin(  x)  .
Adding constant multiples of two solutions gives another solution. Now
1 1
y3 ( x)  y1 ( x)  y2 ( x)
2 2
x
 e cos(  x)
and
i i
y4 ( x)  y2 ( x)  y1 ( x)
2 2
x
 e sin(  x)
which are both real. We can check that functions are indeed solutions to (1) and so we have
found two real solutions to (1) that aren’t constant multiples of each other. Thus, in the case
that the characteristic equation has complex roots the general solution to the DE is

y( x)  Ae x cos( x)  Be x sin(  x) .

Example 5: Find the general solution to y  4 y  13 y  0 .

The characteristic equation for this homogeneous second order linear DE with
constant coefficients is

Page 39
 Differential Equations (IV) - Lecture Notes

r 2  4r  13  0
which has the complex solutions
r  2  3i .
Thus the general solution to the DE is
y( x)  Ae2 x cos(3x)  Be2 x sin(3x) .
☼

The above discussion can be summarised as follows.

To solve the homogeneous second order linear DE with constant coefficients

ay  by  cy  0 (†)

• Find the roots of the characteristic equation

ar 2  br  c  0 .
• If the characteristic equation has two real roots r1 and r2 then the general
solution to (†) is
y( x)  Aer1x  Ber2 x .
• If the characteristic equation has a repeated real root r then the general solution
to (†) is
y( x)  Aerx  Bxerx .
• If the characteristic equation has complex roots r    i then the general
solution to (†) is
y( x)  Ae x cos( x)  Be x sin(  x) .

Given that the general solution to a second order DE has two arbitrary constants in
order to find a particular solution it is necessary to give two conditions. In an initial-value
problem the conditions given are the values of y and y at the same value of x, usually x  0 .
In a boundary-value problem the conditions given are the values of y at two different values
of x. Even when we have a general solution to the DE, it may not always be possible to
satisfy a given set of boundary-value conditions.

Example 6: Solve the initial-value problem

4 y  3 y  y  0 ,
y(0)  3 , y(0)  0 .

The DE is a homogeneous second order linear DE with constant coefficients. The

characteristic equation then is
4r 2  3r  1  0 .
Since this equation has two real solutions,
1
r   ,1
4
the general solution is
y( x)  Ae x /4  Be x .
Thus

Page 40
 Math1120 - Mathematics II

1
y( x)   Ae x /4  Be x .
4
Using the initial conditions gives the equations.
A B  3
1
 A B  0
4 .
Solving these simultaneously gives
12 3
A , B .
5 5

Thus the solution to the initial-value problem is

12 3
y( x)  e x /4  e x .
5 5

1
This solution is shown in Figure 1. Note that since r1    0 the term in the general
4
12  x 4
solution associated with this root (i.e. e ) decays as x   . On the other hand
5
3
since r2  1  0 the term in the general solution associated with this root (i.e. e x )
5
grows as x   .

Figure 1
☼

Page 41
 Differential Equations (IV) - Lecture Notes

Example 7: Solve the boundary-value problem

y  6 y  25 y  0 ,
y(0)  1 , y( / 8)  2 .

The characteristic equation is

r 2  6r  25  0 ,
which has the complex solutions
r  3  4i .
Thus the general solution is
y( x)  e3 x  A cos(4 x)  B sin(4 x)  .

Using the boundary conditions gives the equations

A 1
Be3 /8  2 .

Thus the solution to the boundary-value problem is


y( x)  e3 x cos(4 x)  2e
3 8

sin(4 x) .

This graph of this function is shown in Figure 2. Note that when the characteristic
equation has complex roots, r     i the solutions the homogeneous second order
linear DE with constant coefficients will be periodic in nature with frequency  .
Further, if   0 then the oscillations will grow as x   whereas if   0 the
oscillations will decay as x   . In this case   3 and hence the oscillations increase
in amplitude.

Figure 2
☼

Page 42
 Math1120 - Mathematics II

Example Tasks

ET 1: If x  x(t ) find the solution to the initial-value problem

15x 16x  4x  0
x(0)  1 , x(0)  3 .

ET 2: Find the solution to the boundary-value problem

y  6 y  9 y  0
y(1)  0 , y(2)  3 .

ET 3: Describe the long term behaviour of the solutions to the DE

p(t )  9 p(t ) .

ET 4*: Use the substitution x  e z to solve the DE

d2y dy
x 2 2  3x  2 y  0 , ( x  0) .
dx dx

Page 43
 Differential Equations (IV) - Lecture Notes

Page 44