im03.

qxd 9/21/05 11:04 AM Page 59

CHAPTER 3 Higher Order Linear ODEs

This chapter is new. Its material is a rearranged and somewhat extended version of material
previously contained in some of the sections of Chap 2. The rearrangement is such that
the presentation parallels that in Chap. 2 for second-order ODEs, to facilitate comparisons.

Root Finding
For higher order ODEs you may need Newton’s method or some other method from
Sec. 19.2 (which is independent of other sections in numerics) in work on a calculator
or with your CAS (which may give you a root-finding method directly).

Linear Algebra
The typical student may have taken an elementary linear algebra course simultaneously
with a course on calculus and will know much more than is needed in Chaps. 2 and 3.
Thus Chaps. 7 and 8 need not be taken before Chap. 3.
In particular, although the Wronskian becomes useful in Chap. 3 (whereas for n
2
one hardly needs it), a very modest knowledge of determinants will suffice. (For n
2
and 3, determinants are treated in a reference section, Sec. 7.6.)

SECTION 3.1. Homogeneous Linear ODEs, page 105

Purpose. Extension of the basic concepts and theory in Secs. 2.1 and 2.6 to homogeneous
linear ODEs of any order n. This shows that practically all the essential facts carry over
without change. Linear independence, now more involved as for n
2, causes the
Wronskian to become indispensable (whereas for n
2 it played a marginal role).
Main Content, Important Concepts
Superposition principle for the homogeneous ODE (2)
General solution, basis, particular solution
General solution of (2) with continuous coefficients exists.
Existence and uniqueness of solution of initial value problem (2), (5)
Linear independence of solutions, Wronskian
General solution includes all solutions of (2).
Comment on Order of Material
In Chap. 2 we first gained practical experience and skill and presented the theory of the
homogeneous linear ODE at the end of the discussion, in Sec. 2.6. In this chapter, with
all the experience gained on second-order ODEs, it is more logical to present the whole
theory at the beginning and the solution methods (for linear ODEs with constant
coefficients) afterward. Similarly, the same logic applies to the nonhomogeneous linear
ODE, for which Sec. 3.3 contains the theory as well as the solution methods.

SOLUTIONS TO PROBLEM SET 3.1, page 111

2. Problems 1–5 should give the student a first impression of the changes occurring in
the transition from n
2 to general n.
59
im03.qxd 9/21/05 11:04 AM Page 60

60 Instructor’s Manual

8. Let y1
x  1, y2
x  2, y3
x. Then
y2  2y1  y3
0
shows linear dependence.
10. Linearly independent
12. Linear dependence, since one of the functions is the zero function
14. cos 2x
cos2 x  sin2 x; linearly dependent
16. (x  1)2  (x  1)2  4x
0; linearly dependent
18. Linearly independent
20. Team Project. (a) (1) No. If y1
0, then (4) holds with any k1  0 and the other
kj all zero.
(2) Yes. If S were linearly dependent on I, then (4) would hold with a kj  0 on I,
hence also on J, contradicting the assumption.
(3) Not necessarily. For instance, x 2 and xx are linearly dependent on the interval 0
 x  1, but linearly independent on 1  x  1.
(4) Not necessarily. See the answer to (3).
(5) Yes. See the answer to (2).
(6) Yes. By assumption, k1y1  • • •  kp yp
0 with k1, • • • , kp not all zero (this
refers to the functions in S ), and for T we can add the further functions with coefficients
all zero; then the condition for linear dependence of T is satisfied.
(b) We can use the Wronskian for testing linear independence only if we know that
the given functions are solutions of a homogeneous linear ODE with continuous
coefficients. Other means of testing are the use of functional relations, e.g.,
ln x 2
2 ln x or trigonometric identities, or the evaluation of the given functions at
several values of x, to see whether we can discover proportionality.

SECTION 3.2. Homogeneous Linear ODEs with Constant Coefficients,

page 111
Purpose. Extension of the algebraic solution method for constant-coefficient ODEs from
n
2 (Sec. 2.2) to any n, and discussion of the increased number of possible cases:
Real different roots
Complex simple roots
Real multiple roots
Complex multiple roots
Combinations of the preceding four basic cases
Explanation of these cases in terms of typical examples
Comment on Numerics
In practical cases, one may have to use Newton’s method or another method for computing
(approximate values of) roots in Sec. 19.2.

SOLUTIONS TO PROBLEM SET 3.2, page 115

2. The form of the given functions shows that the characteristic equation has a triple
root 2; hence it is
(  2)3
3  62  12  8
0.
im03.qxd 9/21/05 11:04 AM Page 61

Instructor’s Manual 61

Hence the ODE is

y   6y   12y  8y
0.

4. The first two functions result from a factor 2  1 of the characteristic equation, and
the other two solutions show that the roots i and i are double roots, so that the
characteristic equation is
(2  1)2
0
and the ODE is
yiv  2y   y
0.

6. The given functions show that the characteristic equation is

(  2)(  1)(  1)(  2)
5  53  4
0.
This gives the ODE
yv  5y   4y
0.
8. The characteristic equation is quadratic in
2. The roots are

29
2 292
4
 100

29
2
21
2

y
50/2.
8/2.
The roots of this are 
5 and 2. Hence a general solution is
y
c1e
2x  c2 e 2x  c3 e
5x  c4 e 5x.

10. The characteristic equation is

16(4  _122  _ _1 2
16 )
16(  4 )
0.
1 2

It has the double roots 
_12. This gives the general solution
y
(c1  c2 x)ex/2  (c3  c4 x)e
x/2.

12. The characteristic equation is

4  32  4
(2  1)(2  4)
0.
Its roots are 1 and 2i. Hence the corresponding real general solution is
y
c1e x  c2 e
x  c3 cos 2x  c4 sin 2x.

14. The particular solution, solving the initial value problem, is

y
4e
x  5e
x/2 cos 3x.
From it, the general solution is obvious.
16. y
e x(_5 _
16 cos x  16 sin x)  e
3
x _3
(16 cos x  2_3
16 sin x)

18. y
10e
4x(cos 1.5x  sin 1.5x)  0.05e0.5x

20. Project. (a) Divide the characteristic equation by   1 if e1x is known.
(b) The idea is the same as in Sec. 2.1.
(c) We first produce the standard form because this is the form under which the
equation for z was derived. Division by x 3 gives
3 6 6 1
y  y   ( 2  1)y  ( 3  )y
0.
x x x x
im03.qxd 9/21/05 11:04 AM Page 62

62 Instructor’s Manual

With y1
x, y 1
1, y 1
0, and the coefficients p1 and p2 from the standard equation,

we obtain
3 3 6
xz   [3  ( )x]z  [2( ) • 1  ( 2  1)x]z
0.
x x x
Simplification gives
6 6
xz   (   x)z
x(z   z)
0.
x x
Hence
z
c1e x 

c2 e
x.
By integration we get the answer

y2
x  z dx
(c e
1
x
 c2 e
x  c3)x.

SECTION 3.3. Nonhomogeneous Linear ODEs, page 116

Purpose. To show that the transition from n
2 (Sec. 2.7) to general n introduces no
new ideas but generalizes all results and practical aspects in a straightforward fashion;
this refers to existence, uniqueness, and the need for a particular solution yp to get a general
solution in the form
y
yh  yp.
Comment on Elastic Beams
This is an important standard example of a fourth-order ODE that needs no lengthy
preparations and explanations.
Vibrating beams follow in Problem Set 12.3. This leads to PDEs, since time t comes
in as an additional variable.
Comment on Variation of Parameters
This method looks perhaps more complicated than it is; also the integrals may often be
difficult to evaluate, and handling the higher order determinants that occur may require
some more skill than the average student will have at this time. Thus it may be good to
discuss this matter only rather briefly.

SOLUTIONS TO PROBLEM SET 3.3, page 122

2. The characteristic equation is

3  32  5  39
(2  6  13)(  3)
[(  3)2  4](  3)
0.
Hence a general solution of the homogeneous ODE is
yh
e
3x(c1 cos 2x  c2 sin 2x)  c3 e 3x.
Using the method of undetermined coefficients, substitute yp
a cos x  b sin x,
obtaining
yp
0.7 cos x  0.1 sin x.

4. The characteristic equation is

3  22  5  6
(  1)(  2)(  3)
0.
im03.qxd 9/21/05 11:04 AM Page 63

Instructor’s Manual 63

Hence a general solution of the homogeneous ODE is

yh
c1e
x  c2 e 2x  c3 e
3x.
Since both terms on the right are solutions of the homogeneous ODE, the Modification
Rule for undetermined coefficients applies. Substitution into the nonhomogeneous
ODE gives the particular solution
y
(7  10x)e
3x  (_1  3x)e
x.
p 2

6. The homogeneous Euler–Cauchy equation can be solved as usual by substituting x m.

The auxiliary equation has the roots 2, 0, 1; hence a general solution is
yh
c1x
2  c2  c3 x.
This result can also be obtained by separating variables and integrating twice; that is,
y /y 
4/x, ln y 
4 ln x  c, y 
c1x
4,
and so on. Variation of parameters gives
yp
8e x(x
1  x
2).

8. The characteristic equation of the homogeneous ODE

3  22  9  18
0
has the roots 3, 2, and 3. Hence a general solution of the homogeneous ODE is
yh
c1e
3x  c2 e 2x  c3 e 3x.
This also shows that the function on the right is a solution of the homogeneous ODE.
Hence the Modification Rule applies, and the particular solution obtained is
yp
0.2xe 2x.

10. The characteristic equation of the homogeneous ODE is

4  16
(2  4)(2  4)
0.
Hence a general solution of the homogeneous ODE is
yh
c1e
2x  c2 e 2x  c3 cos 2x  c4 sin 2x.
A particular solution yp of the nonhomogeneous ODE can be obtained by the method
of undetermined coefficients; this yp can be written in terms of exponential or
hyperbolic functions:
yp
1.5(e
2x  e 2x)  2x(e
2x  e 2x)

3 cosh 2x  4x sinh 2x.

Applying the initial conditions, we obtain the answer
y
4e
2x  16 sin 2x  yp,
from which we cannot immediately recognize a basis of solutions. Of course, when
sin 2x occurs, cos 2x must be in the basis. But e 2x remains hidden and cannot be seen
from the answer.
12. The characteristic equation is quadratic in 2, the roots being 5, 1, 1, 5. The right
side requires a quadratic polynomial whose coefficients can be determined by
im03.qxd 9/21/05 11:04 AM Page 64

64 Instructor’s Manual

substitution. Finally, the initial conditions are used to determine the four arbitrary
constants in the general solution of the nonhomogeneous ODE thus obtained. The
answer is
y
5e
x  e
5x  6.16  4x  2x 2.
Again, two of four possible terms resulting from the homogeneous ODE are not visible
in the answer. The student should recognize that all or some or none of the solutions
of a basis of the homogeneous ODE may be present in the final answer; this will
depend on the initial conditions, so the student should experiment a little with this
problem to see what is going on.
14. The method of undetermined coefficients gives
yp
0.08 cos x  0.04 sin x.
A basis of solutions of the homogeneous ODE is e
2x, e 2x, ex/2. From this and the
initial conditions we obtain the answer
y
0.11e
2x  0.15e 2x  0.96ex/2  yp
in which all three basis functions occur.
16. The first equation has as a general solution
y
(c1  c2 x  c3 x 2)e 4x  _8 7/2 4x
105 x e .
Hence in cases such as this, one can try
yp
x1/2(a0  a1x  a2 x 2  a3 x 3)e 4x.
One can now modify the right side systematically and see how the solution changes.
The second ODE has as a general solution
y
c x
2  c x  c x 3  _
1 2 3
1
x(18(ln x)2  6 ln x  7).
216

This shows that undetermined coefficients would not be suitable—the function on the
right gives no clue of what yp may look like.
Of course, the dependence on the left side also remains to be explored.

SOLUTIONS TO CHAP. 3 REVIEW QUESTIONS AND PROBLEMS, page 122

6. The characteristic equation is

3  62  18  40
[(  1)2  9](  4)
0.
Hence a general solution is
y
c1e
4x  e
x(c2 cos 3x  c3 sin 3x).

8. The characteristic equation is quadratic in 2; namely,

(2  1)(2  9)
0.
Hence a general solution is
y
c1 cos x  c2 sin x  c3 cos 3x  c4 sin 3x.

10. The characteristic equation has the triple root 1 because it is

(  1)3
0.
im03.qxd 9/21/05 11:04 AM Page 65

Instructor’s Manual 65

Hence a general solution of the homogeneous ODE is

yh
(c1  c2 x  c3 x 2)e
x.
The method of undetermined coefficients gives the particular solution
yp
x 2  6x  12.

12. The characteristic equation is

2(  2)(  4)
0.
Hence a general solution of the homogeneous ODE is
yh
c1  c2 x  c3 e 4x  c4 e
2x.
The method of undetermined coefficients gives the particular solution of the
nonhomogeneous ODE in the form
yp
0.3 cos 2x  0.1 sin 2x.
14. The auxiliary equation of the homogeneous ODE
x 3y   ax 2y   bxy  cy
0
is
m(m  1)(m  2)  am(m  1)  bm  c

m3  (a  3)m2  (b  a  2)m  c
0.
In our equation, a
3, b
6, and c
6. Accordingly, the auxiliary equation
becomes
m3  6m2  11m  6
(m  1)(m  2)(m  3)
0.
Hence a general solution of the homogeneous ODE is
yh
c1x  c2 x 2  c3 x 3.
Variation of parameters gives the particular solution
yp
0.5x
2.

16. The characteristic equation of the ODE is

3  22  4  8
(  2)(2  4)
0.
Hence a general solution is
yh
c1e 2x  c2 cos 2x  c3 sin 2x.
Using the initial conditions, we obtain the particular solution
y
3e 2x  4 cos 2x  12 sin 2x.
18. The characteristic equation of the homogeneous ODE is
(2  25)
0.
Hence a general solution of the homogeneous ODE is
yh
c1  c2 cos 5x  c3 sin 5x.
im03.qxd 9/21/05 11:04 AM Page 66

66 Instructor’s Manual

Using cos2 x
_12(1  cos 2x), we can apply the method of undetermined coefficients
to obtain a particular solution of the nonhomogeneous ODE in the form
y
1_6 x  _
p 25
2
sin 8x.
39

Finally, from the initial conditions we obtain the answer

224
y
 sin 5x  yp.
4875
20. We can use the formula in the solution of Prob. 14,
m3  (a  3)m2  (b  a  2)m  c
0,
where, in our present ODE, a
5, b
2, and c
2. Hence this formula becomes
m3  2m2  m  2
(m  1)(m  1)(m  2)
0
and gives the basis of solutions x, x
1, x
2. Variation of parameters gives the particular
solution of the nonhomogeneous ODE
yp
1.6x3/2.
From this and the initial conditions we obtain the answer
y
5x
2  4x  1.6x3/2.