Buscar

Chapt_10

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes
Você viu 3, do total de 53 páginas

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes
Você viu 6, do total de 53 páginas

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes
Você viu 9, do total de 53 páginas

Faça como milhares de estudantes: teste grátis o Passei Direto

Esse e outros conteúdos desbloqueados

16 milhões de materiais de várias disciplinas

Impressão de materiais

Agora você pode testar o

Passei Direto grátis

Você também pode ser Premium ajudando estudantes

Prévia do material em texto

Part III Systems of Differential Equations
1010 Systems of LinearDifferential Equations
EXERCISES 10.1
Preliminary Theory
1. Let X =
(
x
y
)
. Then X′ =
(
3 −5
4 8
)
X.
2. Let X =
(
x
y
)
. Then X′ =
(
4 −7
5 0
)
X.
3. Let X =
xy
z
. Then X′ =
−3 4 −96 −1 0
10 4 3
X.
4. Let X =
xy
z
. Then X′ =
 1 −1 01 0 2
−1 0 1
X.
5. Let X =
xy
z
. Then X′ =
 1 −1 12 1 −1
1 1 1
X +
 0−3t2
t2
 +
 t0
−t
 +
−10
2
.
6. Let X =
xy
z
. Then X′ =
−3 4 05 9 0
0 1 6
X +
 e
−t sin 2t
4e−t cos 2t
−e−t
.
7.
dx
dt
= 4x + 2y + et;
dy
dt
= −x + 3y − et
8.
dx
dt
= 7x + 5y − 9z − 8e−2t; dy
dt
= 4x + y + z + 2e5t;
dz
dt
= −2y + 3z + e5t − 3e−2t
9.
dx
dt
= x− y + 2z + e−t − 3t; dy
dt
= 3x− 4y + z + 2e−t + t; dz
dt
= −2x + 5y + 6z + 2e−t − t
10.
dx
dt
= 3x− 7y + 4 sin t + (t− 4)e4t; dy
dt
= x + y + 8 sin t + (2t + 1)e4t
11. Since
X′ =
( −5
−10
)
e−5t and
(
3 −4
4 −7
)
X =
( −5
−10
)
e−5t
551
10.1 Preliminary Theory
we see that
X′ =
(
3 −4
4 −7
)
X.
12. Since
X′ =
(
5 cos t− 5 sin t
2 cos t− 4 sin t
)
et and
(−2 5
−2 4
)
X =
(
5 cos t− 5 sin t
2 cos t− 4 sin t
)
et
we see that
X′ =
(−2 5
−2 4
)
X.
13. Since
X′ =
( 3
2
−3
)
e−3t/2 and
(−1 14
1 −1
)
X =
( 3
2
−3
)
e−3t/2
we see that
X′ =
(−1 1/4
1 −1
)
X.
14. Since
X′ =
(
5
−1
)
et +
(
4
−4
)
tet and
(
2 1
−1 0
)
X =
(
5
−1
)
et +
(
4
−4
)
tet
we see that
X′ =
(
2 1
−1 0
)
X.
15. Since
X′ =
 00
0
 and
 1 2 16 −1 0
−1 −2 −1
X =
 00
0

we see that
X′ =
 1 2 16 −1 0
−1 −2 −1
X.
16. Since
X′ =
 cos t12 sin t− 12 cos t
− cos t− sin t
 and
 1 0 11 1 0
−2 0 −1
X =
 cos t12 sin t− 12 cos t
− cos t− sin t

we see that
X′ =
 1 0 11 1 0
−2 0 −1
X.
17. Yes, since W (X1,X2) = −2e−8t �= 0 the set X1, X2 is linearly independent on −∞ < t < ∞.
18. Yes, since W (X1,X2) = 8e2t �= 0 the set X1, X2 is linearly independent on −∞ < t < ∞.
19. No, since W (X1,X2,X3) = 0 the set X1, X2, X3 is linearly dependent on −∞ < t < ∞.
20. Yes, since W (X1,X2,X3) = −84e−t �= 0 the set X1, X2, X3 is linearly independent on −∞ < t < ∞.
21. Since
X′p =
(
2
−1
)
and
(
1 4
3 2
)
Xp +
(
2
−4
)
t +
( −7
−18
)
=
(
2
−1
)
552
10.1 Preliminary Theory
we see that
X′p =
(
1 4
3 2
)
Xp +
(
2
−4
)
t +
( −7
−18
)
.
22. Since
X′p =
(
0
0
)
and
(
2 1
1 −1
)
Xp +
(−5
2
)
=
(
0
0
)
we see that
X′p =
(
2 1
1 −1
)
Xp +
(−5
2
)
.
23. Since
X′p =
(
2
0
)
et +
(
1
−1
)
tet and
(
2 1
3 4
)
Xp −
(
1
7
)
et =
(
2
0
)
et +
(
1
−1
)
tet
we see that
X′p =
(
2 1
3 4
)
Xp −
(
1
7
)
et.
24. Since
X′p =
 3 cos 3t0
−3 sin 3t
 and
 1 2 3−4 2 0
−6 1 0
Xp +
−14
3
 sin 3t =
 3 cos 3t0
−3 sin 3t

we see that
X′p =
 1 2 3−4 2 0
−6 1 0
Xp +
−14
3
 sin 3t.
25. Let
X1 =
 6−1
−5
 e−t, X2 =
−31
1
 e−2t, X3 =
 21
1
 e3t, and A =
 0 6 01 0 1
1 1 0
 .
Then
X′1 =
−61
5
 e−t = AX1,
X′2 =
 6−2
−2
 e−2t = AX2,
X′3 =
 63
3
 e3t = AX3,
and W (X1,X2,X3) = 20 �= 0 so that X1, X2, and X3 form a fundamental set for X′ = AX on −∞ < t < ∞.
26. Let
X1 =
(
1
−1−√2
)
e
√
2 t,
X2 =
(
1
−1 +√2
)
e−
√
2 t,
553
10.1 Preliminary Theory
Xp =
(
1
0
)
t2 +
(−2
4
)
t +
(
1
0
)
,
and
A =
(−1 −1
−1 1
)
.
Then
X′1 =
( √
2
−2−√2
)
e
√
2 t = AX1,
X′2 =
( −√2
−2 +√2
)
e−
√
2 t = AX2,
X′p =
(
2
0
)
t +
(−2
4
)
= AXp +
(
1
1
)
t2 +
(
4
−6
)
t +
(−1
5
)
,
and W (X1,X2) = 2
√
2 �= 0 so that Xp is a particular solution and X1 and X2 form a fundamental set on
−∞ < t < ∞.
EXERCISES 10.2
Homogeneous Linear Systems
1. The system is
X′ =
(
1 2
4 3
)
X
and det(A− λI) = (λ− 5)(λ + 1) = 0. For λ1 = 5 we obtain(−4 2 0
4 −2 0
)
=⇒
(
1 − 12 0
0 0 0
)
so that K1 =
(
1
2
)
.
For λ2 = −1 we obtain (
2 2 0
4 4 0
)
=⇒
(
1 1 0
0 0 0
)
so that K2 =
(−1
1
)
.
Then
X = c1
(
1
2
)
e5t + c2
(−1
1
)
e−t.
2. The system is
X′ =
(
2 2
1 3
)
X
and det(A− λI) = (λ− 1)(λ− 4) = 0. For λ1 = 1 we obtain(
1 2 0
1 2 0
)
=⇒
(
1 2 0
0 0 0
)
so that K1 =
(−2
1
)
.
For λ2 = 4 we obtain (−2 2 0
1 −1 0
)
=⇒
(−1 1 0
0 0 0
)
so that K2 =
(
1
1
)
.
554
10.2 Homogeneous Linear Systems
Then
X = c1
(−2
1
)
et + c2
(
1
1
)
e4t.
3. The system is
X′ =
( −4 2
− 52 2
)
X
and det(A− λI) = (λ− 1)(λ + 3) = 0. For λ1 = 1 we obtain( −5 2 0
− 52 1 0
)
=⇒
(−5 2 0
0 0 0
)
so that K1 =
(
2
5
)
.
For λ2 = −3 we obtain ( −1 2 0
− 52 5 0
)
=⇒
(−1 2 0
0 0 0
)
so that K2 =
(
2
1
)
.
Then
X = c1
(
2
5
)
et + c2
(
2
1
)
e−3t.
4. The system is
X′ =
(− 52 2
3
4 −2
)
X
and det(A− λI) = 12 (λ + 1)(2λ + 7) = 0. For λ1 = −7/2 we obtain(
1 2 0
3
4
3
2 0
)
=⇒
(
1 2 0
0 0 0
)
so that K1 =
(−2
1
)
.
For λ2 = −1 we obtain (− 32 2 0
3
4 −1 0
)
=⇒
(−3 4 0
0 0 0
)
so that K2 =
(
4
3
)
.
Then
X = c1
(−2
1
)
e−7t/2 + c2
(
4
3
)
e−t.
5. The system is
X′ =
(
10 −5
8 −12
)
X
and det(A− λI) = (λ− 8)(λ + 10) = 0. For λ1 = 8 we obtain(
2 −5 0
8 −20 0
)
=⇒
(
1 − 52 0
0 0 0
)
so that K1 =
(
5
2
)
.
For λ2 = −10 we obtain (
20 −5 0
8 −2 0
)
=⇒
(
1 − 14 0
0 0 0
)
so that K2 =
(
1
4
)
.
Then
X = c1
(
5
2
)
e8t + c2
(
1
4
)
e−10t.
6. The system is
X′ =
(−6 2
−3 1
)
X
555
10.2 Homogeneous Linear Systems
and det(A− λI) = λ(λ + 5) = 0. For λ1 = 0 we obtain(−6 2 0
−3 1 0
)
=⇒
(
1 − 13 0
0 0 0
)
so that K1 =
(
1
3
)
.
For λ2 = −5 we obtain (−1 2 0
−3 6 0
)
=⇒
(
1 −2 0
0 0 0
)
so that K2 =
(
2
1
)
.
Then
X = c1
(
1
3
)
+ c2
(
2
1
)
e−5t.
7. The system is
X′ =
 1 1 −10 2 0
0 1 −1
X
and det(A− λI) = (λ− 1)(2− λ)(λ + 1) = 0. For λ1 = 1, λ2 = 2, and λ3 = −1 we obtain
K1 =
 10
0
 , K2 =
 23
1
 , and K3 =
 10
2
 ,
so that
X = c1
 10
0
 et + c2
 23
1
 e2t + c3
 10
2
 e−t.
8. The system is
X′ =
 2 −7 05 10 4
0 5 2
X
and det(A− λI) = (2− λ)(λ− 5)(λ− 7) = 0. For λ1 = 2, λ2 = 5, and λ3 = 7 we obtain
K1 =
 40
−5
 , K2 =
−73
5
 , and K3 =
−75
5
 ,
so that
X = c1
 40
−5
 e2t + c2
−73
5
 e5t + c3
−75
5
 e7t.
9. We have det(A− λI) = −(λ + 1)(λ− 3)(λ + 2) = 0. For λ1 = −1, λ2 = 3, and λ3 = −2 we obtain
K1 =
−10
1
 , K2 =
 14
3
 , and K3 =
 1−1
3
 ,
so that
X = c1
−10
1
 e−t + c2
 14
3
 e3t + c3
 1−1
3
 e−2t.
556
10.2 Homogeneous Linear Systems
10. We have det(A− λI) = −λ(λ− 1)(λ− 2) = 0. For λ1 = 0, λ2 = 1, and λ3 = 2 we obtain
K1 =
 10
−1
 , K2 =
 01
0
 , and K3 =
 10
1
 ,
so that
X = c1
 10
−1
 + c2 01
0
 et + c3
 10
1
 e2t.
11. We have det(A− λI) = −(λ + 1)(λ + 1/2)(λ + 3/2) = 0. For λ1 = −1, λ2 = −1/2, and λ3 = −3/2 we obtain
K1 =
 40
−1
 , K2 =
−126
5
 , and K3 =
 42
−1
 ,
so that
X = c1
 40
−1
 e−t + c2
−126
5
 e−t/2 + c3
 42
−1
 e−3t/2.
12. We have det(A− λI) = (λ− 3)(λ + 5)(6− λ) = 0. For λ1 = 3, λ2 = −5, and λ3 = 6 we obtain
K1 =
 11
0
 , K2 =
 1−1
0
 , and K3 =
 2−2
11
 ,
so that
X = c1
 11
0
 e3t + c2
 1−1
0
 e−5t + c3
 2−2
11
 e6t.
13. We have det(A− λI) = (λ + 1/2)(λ− 1/2) = 0. For λ1 = −1/2 and λ2 = 1/2 we obtain
K1 =
(
0
1
)
and K2 =
(
1
1
)
,
so that
X = c1
(
0
1
)
e−t/2 + c2
(
1
1
)
et/2.
If
X(0) =
(
3
5
)
then c1 = 2 and c2 = 3.
14. We have det(A− λI) = (2− λ)(λ− 3)(λ + 1) = 0. For λ1 = 2, λ2 = 3, and λ3 = −1 we obtain
K1 =
 5−3
2
 , K2 =
 20
1
 , and K3 =
−20
1
 ,
so that
X = c1
 5−3
2
 e2t + c2
 20
1
 e3t + c3
−20
1
 e−t.
557
10.2 Homogeneous Linear Systems
If
X(0) =
 13
0

then c1 = −1, c2 = 5/2, and c3 = −1/2.
15. X = c1
 0.3821750.851161
0.359815
 e8.58979t + c2
 0.405188−0.676043
0.615458
 e2.25684t + c3
−0.923562−0.132174
0.35995
 e−0.0466321t
16. X = c1

0.0312209
0.949058
0.239535
0.195825
0.0508861
 e5.05452t + c2

−0.280232
−0.836611
−0.275304
0.176045
0.338775
 e4.09561t + c3

0.262219
−0.162664
−0.826218
−0.346439
0.31957
 e−2.92362t
+c4

0.313235
0.64181
0.31754
0.173787
−0.599108
 e2.02882t + c5

−0.301294
0.466599
0.222136
0.0534311
−0.799567
 e−0.155338t
17. (a)
(b) Letting c1 = 1 and c2 = 0 we get x = 5e8t, y = 2e8t. Eliminating the parameter we find y = 25x, x > 0.
When c1 = −1 and c2 = 0 we find y = 25x, x < 0. Letting c1 = 0 and c2 = 1 we get x = e−10t, y = 4e−10t.
Eliminating the parameter we find y = 4x, x > 0. Letting c1 = 0 and c2 = −1 we find y = 4x, x < 0.
(c) The eigenvectors K1 = (5, 2) and K2 = (1, 4) are shown in the figure in part (a).
18. In Problem 2, letting c1 = 1 and c2 = 0 we get x = −2et, y = et.
Eliminating the parameter we find y = − 12x, x < 0. When c1 = −1
and c2 = 0 we find y = − 12x, x > 0. Letting c1 = 0 and c2 = 1 we
get x = e4t, y = e4t. Eliminating the parameter we find y = x, x > 0.
When c1 = 0 and c2 = −1 we find y = x, x < 0.
558
10.2 Homogeneous Linear Systems
In Problem 4, letting c1 = 1 and c2 = 0 we get x = −2e−7t/2, y =
e−7t/2. Eliminating the parameter we find y = − 12x, x < 0. When
c1 = −1 and c2 = 0 we find y = − 12x, x > 0. Letting c1 = 0 and
c2 = 1 we get x = 4e−t, y = 3e−t. Eliminating the parameter we find
y = 34x, x > 0. When c1 = 0 and c2 = −1 we find y = 34x, x < 0.
19. We have det(A− λI) = λ2 = 0. For λ1 = 0 we obtain
K =
(
1
3
)
.
A solution of (A− λ1I)P = K is
P =
(
1
2
)
so that
X = c1
(
1
3
)
+ c2
[(
1
3
)
t +
(
1
2
)]
.
20. We have det(A− λI) = (λ + 1)2 = 0. For λ1 = −1 we obtain
K =
(
1
1
)
.
A solution of (A− λ1I)P = K is
P =
(
0
1
5
)
so that
X = c1
(
1
1
)
e−t + c2
[(
1
1
)
te−t +
(
0
1
5
)
e−t
]
.
21. We have det(A− λI) = (λ− 2)2 = 0. For λ1 = 2 we obtain
K =
(
1
1
)
.
A solution of (A− λ1I)P = K is
P =
(− 13
0
)
so that
X = c1
(
1
1
)
e2t + c2
[(
1
1
)
te2t +
(− 13
0
)
e2t
]
.
22. We have det(A− λI) = (λ− 6)2 = 0. For λ1 = 6 we obtain
K =
(
3
2
)
.
A solution of (A− λ1I)P = K is
P =
( 1
2
0
)
559
10.2 Homogeneous Linear Systems
so that
X = c1
(
3
2
)
e6t + c2
[(
3
2
)
te6t +
( 1
2
0
)
e6t
]
.
23. We have det(A− λI) = (1− λ)(λ− 2)2 = 0. For λ1 = 1 we obtain
K1 =
 11
1
 .
For λ2 = 2 we obtain
K2 =
 10
1
 and K3 =
 11
0
 .
Then
X = c1
 11
1
 et + c2
 10
1
 e2t + c3
 11
0
 e2t.
24. We have det(A− λI) = (λ− 8)(λ + 1)2 = 0. For λ1 = 8 we obtain
K1 =
 21
2
 .
For λ2 = −1 we obtain
K2 =
 0−2
1
 and K3 =
 1−2
0
 .
Then
X = c1
 21
2
 e8t + c2
 0−2
1
 e−t + c3
 1−2
0
 e−t.
25. We have det(A− λI) = −λ(5− λ)2 = 0. For λ1 = 0 we obtain
K1 =
−4−5
2
 .
For λ2 = 5 we obtain
K =
−20
1
 .
A solution of (A− λ1I)P = K is
P =

5
2
1
2
0

560
10.2 Homogeneous Linear Systems
so that
X = c1
−4−5
2
 + c2
−20
1
 e5t + c3

−20
1
 te5t +

5
2
1
2
0
 e5t
 .
26. We have det(A− λI) = (1− λ)(λ− 2)2 = 0. For λ1 = 1 we obtain
K1 =
 10
0
 .
For λ2 = 2 we obtain
K =
 0−1
1
 .
A solution of (A− λ2I)P = K is
P =
 0−1
0

so that
X = c1
 10
0
 et + c2
 0−1
1
 e2t + c3

 0−1
1
 te2t +
 0−1
0
 e2t
 .
27. We have det(A− λI) = −(λ− 1)3 = 0. For λ1 = 1 we obtain
K =
 01
1
 .
Solutions of (A− λ1I)P = K and (A− λ1I)Q = P are
P =
 01
0
 and Q =

1
2
0
0

so that
X = c1
 01
1
 et + c2

 01
1
 tet +
 01
0
 et
 + c3

 01
1
 t22 et +
 01
0
 tet +

1
2
0
0
 et
 .
28. We have det(A− λI) = (λ− 4)3 = 0. For λ1 = 4 we obtain
K =
 10
0
 .
Solutions of (A− λ1I)P = K and (A− λ1I)Q = P are
P =
 01
0
 and Q =
 00
1

561
10.2 Homogeneous Linear Systems
so that
X = c1
 10
0
 e4t + c2

 10
0
 te4t +
 01
0
 e4t
 + c3

 10
0
 t22 e4t +
 01
0
 te4t +
 00
1
 e4t
 .
29. We have det(A− λI) = (λ− 4)2 = 0. For λ1 = 4 we obtain
K =
(
2
1
)
.
A solution of (A− λ1I)P = K is
P =
(
1
1
)
so that
X = c1
(
2
1
)
e4t + c2
[(
2
1
)
te4t +
(
1
1
)
e4t
]
.
If
X(0) =
(−1
6
)
then c1 = −7 and c2 = 13.
30. We have det(A− λI) = −(λ + 1)(λ− 1)2 = 0. For λ1 = −1 we obtain
K1 =
−10
1
 .
For λ2 = 1 we obtain
K2 =
 10
1
 and K3 =
 01
0

so that
X = c1
−10
1
 e−t + c2
 10
1
 et + c3
 01
0
 et.
If
X(0) =
 12
5

then c1 = 2, c2 = 3, and c3 = 2.
31. In this case det(A − λI) = (2 − λ)5, and λ1 = 2 is an eigenvalue of multiplicity 5. Linearly independent
eigenvectors are
K1 =

1
0
0
0
0
 , K2 =

0
0
1
0
0
 , and K3 =

0
0
0
1
0
 .
32. In Problem 20 letting c1 = 1 and c2 = 0 we get x = et, y = et. Eliminating the parameter we find y = x, x > 0.
When c1 = −1 and c2 = 0 we find y = x, x < 0.
562
10.2 Homogeneous Linear Systems
In Problem 21 letting c1 = 1 and c2 = 0 we get x = e2t, y = e2t. Eliminating the parameter we find y = x,
x > 0. When c1 = −1 and c2 = 0 we find y = x, x < 0.
In Problems 33-46 the form of the answer will vary according to the choice of eigenvector. For example, in
Problem 33, if K1 is chosen to be
(
1
2− i
)
the solution has the form
X = c1
(
cos t
2 cos t + sin t
)
e4t + c2
(
sin t
2 sin t− cos t
)
e4t.
33. We have det(A− λI) = λ2 − 8λ + 17 = 0. For λ1 = 4 + i we obtain
K1 =
(
2 + i
5
)
so that
X1 =
(
2 + i
5
)
e(4+i)t =
(
2 cos t− sin t
5 cos t
)
e4t + i
(
cost + 2 sin t
5 sin t
)
e4t.
Then
X = c1
(
2 cos t− sin t
5 cos t
)
e4t + c2
(
cos t + 2 sin t
5 sin t
)
e4t.
34. We have det(A− λI) = λ2 + 1 = 0. For λ1 = i we obtain
K1 =
(−1− i
2
)
so that
X1 =
(−1− i
2
)
eit =
(
sin t− cos t
2 cos t
)
+ i
(− cos t− sin t
2 sin t
)
.
Then
X = c1
(
sin t− cos t
2 cos t
)
+ c2
(− cos t− sin t
2 sin t
)
.
35. We have det(A− λI) = λ2 − 8λ + 17 = 0. For λ1 = 4 + i we obtain
K1 =
(−1− i
2
)
so that
X1 =
(−1− i
2
)
e(4+i)t =
(
sin t− cos t
2 cos t
)
e4t + i
(− sin t− cos t
2 sin t
)
e4t.
563
10.2 Homogeneous Linear Systems
Then
X = c1
(
sin t− cos t
2 cos t
)
e4t + c2
(− sin t− cos t
2 sin t
)
e4t.
36. We have det(A− λI) = λ2 − 10λ + 34 = 0. For λ1 = 5 + 3i we obtain
K1 =
(
1− 3i
2
)
so that
X1 =
(
1− 3i
2
)
e(5+3i)t =
(
cos 3t + 3 sin 3t
2 cos 3t
)
e5t + i
(
sin 3t− 3 cos 3t
2 sin 3t
)
e5t.
Then
X = c1
(
cos 3t + 3 sin 3t
2 cos 3t
)
e5t + c2
(
sin 3t− 3 cos 3t
2 sin 3t
)
e5t.
37. We have det(A− λI) = λ2 + 9 = 0. For λ1 = 3i we obtain
K1 =
(
4 + 3i
5
)
so that
X1 =
(
4 + 3i
5
)
e3it =
(
4 cos 3t− 3 sin 3t
5 cos 3t
)
+ i
(
4 sin 3t + 3 cos 3t
5 sin 3t
)
.
Then
X = c1
(
4 cos 3t− 3 sin 3t
5 cos 3t
)
+ c2
(
4 sin 3t + 3 cos 3t
5 sin 3t
)
.
38. We have det(A− λI) = λ2 + 2λ + 5 = 0. For λ1 = −1 + 2i we obtain
K1 =
(
2 + 2i
1
)
so that
X1 =
(
2 + 2i
1
)
e(−1+2i)t =
(
2 cos 2t− 2 sin 2t
cos 2t
)
e−t + i
(
2 cos 2t + 2 sin 2t
sin 2t
)
e−t.
Then
X = c1
(
2 cos 2t− 2 sin 2t
cos 2t
)
e−t + c2
(
2 cos 2t + 2 sin 2t
sin 2t
)
e−t.
39. We have det(A− λI) = −λ (λ2 + 1) = 0. For λ1 = 0 we obtain
K1 =
 10
0
 .
For λ2 = i we obtain
K2 =
−ii
1

so that
X2 =
−ii
1
 eit =
 sin t− sin t
cos t
 + i
− cos tcos t
sin t
 .
564
10.2 Homogeneous Linear Systems
Then
X = c1
 10
0
 + c2
 sin t− sin t
cos t
 + c3
− cos tcos t
sin t
 .
40. We have det(A− λI) = −(λ + 3)(λ2 − 2λ + 5) = 0. For λ1 = −3 we obtain
K1 =
 0−2
1
 .
For λ2 = 1 + 2i we obtain
K2 =
−2− i−3i
2

so that
X2 =
−2 cos 2t + sin 2t3 sin 2t
2 cos 2t
 et + i
− cos 2t− 2 sin 2t−3 cos 2t
2 sin 2t
 et.
Then
X = c1
 0−2
1
 e−3t + c2
−2 cos 2t + sin 2t3 sin 2t
2 cos 2t
 et + c3
− cos 2t− 2 sin 2t−3 cos 2t
2 sin 2t
 et.
41. We have det(A− λI) = (1− λ)(λ2 − 2λ + 2) = 0. For λ1 = 1 we obtain
K1 =
 02
1
 .
For λ2 = 1 + i we obtain
K2 =
 1i
i

so that
X2 =
 1i
i
 e(1+i)t =
 cos t− sin t
− sin t
 et + i
 sin tcos t
cos t
 et.
Then
X = c1
 02
1
 et + c2
 cos t− sin t
− sin t
 et + c3
 sin tcos t
cos t
 et.
42. We have det(A− λI) = −(λ− 6)(λ2 − 8λ + 20) = 0. For λ1 = 6 we obtain
K1 =
 01
0
 .
565
10.2 Homogeneous Linear Systems
For λ2 = 4 + 2i we obtain
K2 =
−i0
2

so that
X2 =
−i0
2
 e(4+2i)t =
 sin 2t0
2 cos 2t
 e4t + i
− cos 2t0
2 sin 2t
 e4t.
Then
X = c1
 01
0
 e6t + c2
 sin 2t0
2 cos 2t
 e4t + c3
− cos 2t0
2 sin 2t
 e4t.
43. We have det(A− λI) = (2− λ)(λ2 + 4λ + 13) = 0. For λ1 = 2 we obtain
K1 =
 28−5
25
 .
For λ2 = −2 + 3i we obtain
K2 =
 4 + 3i−5
0

so that
X2 =
 4 + 3i−5
0
 e(−2+3i)t =
 4 cos 3t− 3 sin 3t−5 cos 3t
0
 e−2t + i
 4 sin 3t + 3 cos 3t−5 sin 3t
0
 e−2t.
Then
X = c1
 28−5
25
 e2t + c2
 4 cos 3t− 3 sin 3t−5 cos 3t
0
 e−2t + c3
 4 sin 3t + 3 cos 3t−5 sin 3t
0
 e−2t.
44. We have det(A− λI) = −(λ + 2)(λ2 + 4) = 0. For λ1 = −2 we obtain
K1 =
 0−1
1
 .
For λ2 = 2i we obtain
K2 =
−2− 2i1
1

so that
X2 =
−2− 2i1
1
 e2it =
−2 cos 2t + 2 sin 2tcos 2t
cos 2t
 + i
−2 cos 2t− 2 sin 2tsin 2t
sin 2t
 .
566
10.2 Homogeneous Linear Systems
Then
X = c1
 0−1
1
 e−2t + c2
−2 cos 2t + 2 sin 2tcos 2t
cos 2t
 + c3
−2 cos 2t− 2 sin 2tsin 2t
sin 2t
 .
45. We have det(A− λI) = (1− λ)(λ2 + 25) = 0. For λ1 = 1 we obtain
K1 =
 25−7
6
 .
For λ2 = 5i we obtain
K2 =
 1 + 5i1
1

so that
X2 =
 1 + 5i1
1
 e5it =
 cos 5t− 5 sin 5tcos 5t
cos 5t
 + i
 sin 5t + 5 cos 5tsin 5t
sin 5t
 .
Then
X = c1
 25−7
6
 et + c2
 cos 5t− 5 sin 5tcos 5t
cos 5t
 + c3
 sin 5t + 5 cos 5tsin 5t
sin 5t
 .
If
X(0) =
 46
−7

then c1 = c2 = −1 and c3 = 6.
46. We have det(A− λI) = λ2 − 10λ + 29 = 0. For λ1 = 5 + 2i we obtain
K1 =
(
1
1− 2i
)
so that
X1 =
(
1
1− 2i
)
e(5+2i)t =
(
cos 2t
cos 2t + 2 sin 2t
)
e5t + i
(
sin 2t
sin 2t− 2 cos 2t
)
e5t.
and
X = c1
(
cos 2t
cos 2t + 2 sin 2t
)
e5t + c3
(
sin 2t
sin 2t− 2 cos 2t
)
e5t.
If X(0) =
(−2
8
)
, then c1 = −2 and c2 = 5.
567
10.2 Homogeneous Linear Systems
47.
48. (a) From det(A− λI) = λ(λ− 2) = 0 we get λ1 = 0 and λ2 = 2. For λ1 = 0 we obtain(
1 1 0
1 1 0
)
=⇒
(
1 1 0
0 0 0
)
so that K1 =
(−1
1
)
.
For λ2 = 2 we obtain(−1 1 0
1 −1 0
)
=⇒
(−1 1 0
0 0 0
)
so that K2 =
(
1
1
)
.
Then
X = c1
(−1
1
)
+ c2
(
1
1
)
e2t.
The line y = −x is not a trajectory of the
system. Trajectories are x = −c1 + c2e2t,
y = c1 +c2e2t or y = x+2c1. This is a family
of lines perpendicular to the line y = −x. All
of the constant solutions of the system do,
however, lie on the line y = −x.
(b) From det(A− λI) = λ2 = 0 we get λ1 = 0 and
K =
(−1
1
)
.
A solution of (A− λ1I)P = K is
P =
(−1
0
)
so that
X = c1
(−1
1
)
+ c2
[(−1
1
)
t +
(−1
0
)]
.
All trajectories are parallel to y = −x, but
y = −x is not a trajectory. There are con-
stant solutions of the system, however, that
do lie on the line y = −x.
568
10.2 Homogeneous Linear Systems
49. The system of differential equations is
x′1 = 2x1 + x2
x′2 = 2x2
x′3 = 2x3
x′4 = 2x4 + x5
x′5 = 2x5.
We see immediately that x2 = c2e2t, x3 = c3e2t, and x5 = c5e2t. Then
x′1 = 2x1 + c2e
2t so x1 = c2te2t + c1e2t,
and
x′4 = 2x4 + c5e
2t so x4 = c5te2t + c4e2t.
The general solution of the system is
X =

c2te
2t + c1e2t
c2e
2t
c3e
2t
c5te
2t + c4e2t
c5e
2t

= c1

1
0
0
0
0
 e2t + c2


1
0
0
0
0
 te2t +

0
1
0
0
0
 e2t

+ c3

0
0
1
0
0
 e2t + c4

0
0
0
1
0
 e2t + c5


0
0
0
1
0
 te2t +

0
0
0
0
1
 e2t

= c1K1e2t + c2
K1te2t +

0
1
0
0
0
 e2t

+ c3K2e2t + c4K3e2t + c5
K3te2t +

0
0
0
0
1
 e2t
 .
There are three solutions of the form X = Ke2t, where K is an eigenvector, and two solutions of the form
X = Kte2t + Pe2t. See (12) in the text. From (13) and (14) in the text
(A− 2I)K1 = 0
and
569
10.2 Homogeneous Linear Systems
(A− 2I)K2 = K1.
This implies 
0 1 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 1
0 0 0 0 0


p1
p2
p3
p4
p5
 =

1
0
0
0
0
 ,
so p2 = 1 and p5 = 0, whilep1, p3, and p4 are arbitrary. Choosing p1 = p3 = p4 = 0 we have
P =

1
0
0
0
0
 .
Therefore a solution is
X =

1
0
0
0
0
 te2t +

1
0
0
0
0
 e2t.
Repeating for K3 we find
P =

0
0
0
0
1
 ,
so another solution is
X =

0
0
0
1
0
 te2t +

0
0
0
0
1
 e2t.
50. From x = 2 cos 2t− 2 sin 2t, y = − cos 2t we find x + 2y = −2 sin 2t. Then
(x + 2y)2 = 4 sin2 2t = 4(1− cos2 2t) = 4− 4 cos2 2t = 4− 4y2
and
x2 + 4xy + 4y2 = 4− 4y2 or x2 + 4xy + 8y2 = 4.
This is a rotated conic section and, from the discriminant b2 − 4ac = 16 − 32 < 0, we see that the curve is an
ellipse.
51. Suppose the eigenvalues are α ± iβ, β > 0. In Problem 36 the eigenvalues are 5 ± 3i, in Problem 37 they are
±3i, and in Problem 38 they are −1± 2i. From Problem 47 we deduce that the phase portrait will consist of a
family of closed curves when α = 0 and spirals when α �= 0. The origin will be a repellor when α > 0, and an
attractor when α < 0.
570
10.3 Solution by Diagonalization
52. (a) The given system can be written as
x′′1 = −
k1 + k2
m1
x1 +
k2
m1
x2, x
′′
2 =
k2
m2
x1 − k2
m2
x2.
In terms of matrices this is X′′ = AX where
X =
(
x1
x2
)
and A =
−
k1 + k2
m1
k2
m1
k2
m2
− k2
m2
 .
(b) If X = Keωt then X′′ = ω2Keωt and AX = AKeωt so that X′′ = AX becomes ω2Keωt = AKeωt or
(A− ω2I)K = 0. Now let ω2 = λ.
(c) When m1 = 1, m2 = 1, k1 = 3, and k2 = 2 we obtain A =
(−5 2
2 −2
)
. The eigenvalues and corresponding
eigenvectors of A are λ1 = −1, λ2 = −6, K1 =
(
1
2
)
, K2 =
(−2
1
)
. Since ω1 = i, ω2 = −i, ω3 =
√
6 i, and
ω4 = −
√
6 i a solution is
X = c1
(
1
2
)
eit + c2
(
1
2
)
e−it + c3
(−2
1
)
e
√
6 it + c4
(−2
1
)
e−
√
6 it.
(d) Using eit = cos t + i sin t and e
√
6 it = cos
√
6 t + i sin
√
6 t the preceding solution can be rewritten as
X = c1
(
1
2
)
(cos t + i sin t) + c2
(
1
2
)
(cos t− i sin t)
+ c3
(−2
1
)
(cos
√
6 t + i sin
√
6 t) + c4
(−2
1
)
(cos
√
6 t + i sin
√
6 t)
= (c1 + c2)
(
1
2
)
cos t + i(c1 − c2)
(
1
2
)
sin t
+ (c3 + c4)
(−2
1
)
cos
√
6 t + i(c3 − c4)
(−2
1
)
sin
√
6 t
= b1
(
1
2
)
cos t + b2
(
1
2
)
sin t + b3
(−2
1
)
cos
√
6 t + b4
(−2
1
)
sin
√
6 t
where b1 = c1 + c2, b2 = i(c1 − c2), b3 = c3 + c4, and b4 = i(c3 − c4).
EXERCISES 10.3
Solution by Diagonalization
1. λ1 = 7, λ2 = −4, K1 =
(
3
1
)
, K2 =
(−2
3
)
, P =
(
3 −2
1 3
)
;
X = PY =
(
3 −2
1 3
) (
c1e
7t
c2e
−4t
)
=
(
3c1e7t − 2c2e−4t
c1e
7t + 3c2e−4t
)
= c1
(
3
1
)
e7t + c2
(−2
3
)
e−4t
571
10.3 Solution by Diagonalization
2. λ1 = 0, λ2 = 1, K1 =
(
1
−1
)
, K2 =
(
1
1
)
, P =
(
1 1
−1 1
)
;
X = PY =
(
1 1
−1 1
) (
c1
c2e
t
)
=
(
c1 + c2et
−c1 + c2et
)
= c1
(
1
−1
)
+ c2
(
1
1
)
et
3. λ1 = 12 , λ2 =
3
2 , K1 =
(
1
−2
)
, K2 =
(
1
2
)
, P =
(
1 1
−2 2
)
;
X = PY =
(
1 1
−2 2
) (
c1e
t/2
c2e
3t/2
)
=
(
c1e
t/2 + c2e3t/2
−2c1et/2 + 2c2e3t/2
)
= c1
(
1
−2
)
et/2 + c2
(
1
2
)
e3t/2
4. λ1 = −
√
2 , λ2 =
√
2 , K1 =
( −1
1 +
√
2
)
, K2 =
( −1
1−√2
)
, P =
( −1 −1
1 +
√
2 1−√2
)
;
X = PY =
( −1 −1
1 +
√
2 1−√2
) (
c1e
−√2 t
c2e
√
2 t
)
=
( −c1e−√2 t − c2e√2 t
(1 +
√
2 )c1e−
√
2 t + (1−√2 )c2e
√
2 t
)
= c1
( −1
1 +
√
2
)
e−
√
2 t + c2
( −1
1−√2
)
e
√
2 t
5. λ1 = −4, λ2 = 2, λ3 = 6, K1 =
−11
0
, K2 =
 11
1
, K3 =
 00
1
, P =
−1 1 01 1 0
0 1 1
;
X = PY =
−1 1 01 1 0
0 1 1

 c1e
−4t
c2e
2t
c3e
6t
 =
−c1e
−4t + c2e2t
c1e
−4t + c2e2t
c2e
2t + c3e6t
 = c1
−11
0
 e−4t + c2
 11
0
 e2t + c3
 00
1
 e6t
6. λ1 = −1, λ2 = 1, λ3 = 4, K1 =
−10
1
, K2 =
 1−2
1
, K3 =
 11
1
, P =
−1 1 10 −2 1
1 1 1
;
X = PY =
−1 1 10 −2 1
1 1 1

 c1e
−t
c2e
t
c3e
4t
 =
 −c1 + c2e
t + c3e4t
−2c2et + c3e4t
c1e
−t + c2et + c3e4t
 = c1
−10
1
 e−t+c2
 1−2
1
 et+c3
 11
1
 e4t
7. λ1 = −1, λ2 = 2, λ3 = 2, K1 =
 11
1
, K2 =
−11
0
, K3 =
−10
1
, P =
 1 −1 −11 1 0
1 0 1
;
X = PY =
 1 −1 −11 1 0
1 0 1

 c1e
−t
c2e
2t
c3e
2t
 =
 c1e
−t − c2e2t − c3e2t
c1e
−t + c2e2t
c1e
−t + c3e2t

= c1
 11
1
 e−t + c2
−11
0
 e2t + c3
−10
1
 e2t
572
10.3 Solution by Diagonalization
8. λ1 = 0, λ2 = 0, λ3 = 0, λ4 = 4, K1 =

−1
1
0
0
, K2 =

−1
0
1
0
, K3 =

−1
0
0
1
, K4 =

1
1
1
1
,
P =

−1 −1 −1 1
1 0 0 1
0 1 0 1
0 0 1 1
;
X = PY =

−1 −1 −1 1
1 0 0 1
0 1 0 1
0 0 1 1


c1
c2
c3
c4e
4t
 =

−c1 − c2 − c3 + c4e4t
c1 + c4e4t
c2 + c4e4t
c3 + c4e4t

= c1

−1
1
0
0
 + c2

−1
0
1
0
 + c3

−1
0
0
1
 + c4

1
1
1
1
 e4t
9. λ1 = 1, λ2 = 2, λ3 = 3, K1 =
 11
1
, K2 =
 22
3
, K3 =
 34
5
, P =
 1 2 31 2 4
1 3 5
;
X = PY =
 1 2 31 2 4
1 3 5

 c1e
t
c2e
2t
c3e
3t
 =
 c1e
t + 2c2e2t + 3c3e3t
c1e
t + 2c2e2t + 4c3e3t
c1e
t + 3c2e2t + 5c3e3t
 = c1
 11
1
 et + c2
 22
3
 e2t + c3
 34
5
 e3t
10. λ1 = 0, λ2 = −2
√
2 , λ3 = 2
√
2 , K1 =
 10
−1
, K2 =
 1−√2
1
, K3 =
 1√2
1
, P =
 1 1 10 −√2 √2
−1 1 1
;
X = PY =
 1 1 10 −√2 √2
−1 1 1

 c1c2e−2√2 t
c3e
2
√
2 t
 =
 c1 + c2e
−2√2 t + c3e2
√
2 t
−√2 c2e−2
√
2 t +
√
2 c3e2
√
2 t
−c1 + c2e−2
√
2 t + c3e2
√
2 t

= c1
 10
−1
 + c2
 1−√2
1
 e−2√2 t + c3
 1√2
1
 e2√2 t
11. (a) Since M =
(
m1 0
0 m2
)
is a diagonal matrix with nonzero diagonal entries, it has an inverse. Writing the
system in the form
m1x
′′
1 + (k1 + k2)x1 − k2x2 = 0
m2x
′′
2 − k2x1 + k2x2 = 0
we see that K =
(
k1 + k2 −k2
−k2 k2
)
.
(b) Since M has an inverse, MX′′ + KX = 0 can be written as X′′ + M−1KX = 0 or X′′ + BX = 0 where
573
10.3 Solution by Diagonalization
B = M−1K =
 1m1 0
0
1
m2
 ( k1 + k2 −k2
−k2 k2
)
=

k1 + k2
m1
− k2
m1
− k2
m2
k2
m2
 .
(c) With m1 = 1,m2 = 1, k1 = 3, and k2 = 2 we have B =
(
5 −2
−2 2
)
. The eigenvalues of B are λ1 = 1 and
λ2 = 6 with corresponding eigenvectors
(
1
2
)
and
(−2
1
)
. Letting X = PY the system can be written
PY′′ + BPY = 0 or Y′′ + P−1BPY = 0 where
(
1 −2
2 1
)
and P−1BP =
(
1 0
0 6
)
. The system is then
Y′′ +
(
1 0
0 6
)
Y = 0, which is uncoupled and equivalent to y′′1 + y1 = 0 and y
′′
2 + 6y2 = 0. The solutions
are y1 = c1 cos t + c2 sin t and y2 = c3 cos
√
6 t + c4 sin
√
6 t.
(d) From
X = PY =
(
1 −2
2 1
) (
y1
y2
)
=
(
y1 − 2y2
2y1 + y2
)
we have
x1 = c1 cos t + c2 sin t− 2c3 cos
√
6 t− 2c4 sin
√
6 t
x1 = 2c1 cos t + 2c2 sin t + c3cos
√
6 t + c4 sin
√
6 t
which is the same as
X = c1
(
1
2
)
cos t + c2
(
1
2
)
sin t + c3
(−2
1
)
cos
√
6 t + c4
(−2
1
)
sin
√
6 t.
EXERCISES 10.4
Nonhomogeneous Linear Systems
1. Solving
det(A− λI) =
∣∣∣∣ 2− λ 3−1 −2− λ
∣∣∣∣ = λ2 − 1 = (λ− 1)(λ + 1) = 0
we obtain eigenvalues λ1 = −1 and λ2 = 1. Corresponding eigenvectors are
K1 =
(−1
1
)
and K2 =
(−3
1
)
.
Thus
Xc = c1
(−1
1
)
e−t + c2
(−3
1
)
et.
Substituting
Xp =
(
a1
b1
)
574
10.4 Nonhomogeneous Linear Systems
into the system yields
2a1 + 3b1 = 7
−a1 − 2b1 = −5,
from which we obtain a1 = −1 and b1 = 3. Then
X(t) = c1
(−1
1
)
e−t + c2
(−3
1
)
et +
(−1
3
)
.
2. Solving
det(A− λI) =
∣∣∣∣ 5− λ 9−1 11− λ
∣∣∣∣ = λ2 − 16λ + 64 = (λ− 8)2 = 0
we obtain the eigenvalue λ = 8. A corresponding eigenvector is
K =
(
3
1
)
.
Solving (A− 8I)P = K we obtain
P =
(
2
1
)
.
Thus
Xc = c1
(
3
1
)
e8t + c2
[(
3
1
)
te8t +
(
2
1
)
e8t
]
.
Substituting
Xp =
(
a1
b1
)
into the system yields
5a1 + 9b1 = −2
−a1 + 11b1 = −6,
from which we obtain a1 = 1/2 and b1 = −1/2. Then
X(t) = c1
(
3
1
)
e8t + c2
[(
3
1
)
te8t +
(
2
1
)
e8t
]
+
( 1
2
− 12
)
.
3. Solving
det(A− λI) =
∣∣∣∣ 1− λ 33 1− λ
∣∣∣∣ = λ2 − 2λ− 8 = (λ− 4)(λ + 2) = 0
we obtain eigenvalues λ1 = −2 and λ2 = 4. Corresponding eigenvectors are
K1 =
(
1
−1
)
and K2 =
(
1
1
)
.
Thus
Xc = c1
(
1
−1
)
e−2t + c2
(
1
1
)
e4t.
Substituting
Xp =
(
a3
b3
)
t2 +
(
a2
b2
)
t +
(
a1
b1
)
into the system yields
a3 + 3b3 = 2
3a3 + b3 = 0
a2 + 3b2 = 2a3
3a2 + b2 + 1 = 2b3
a1 + 3b1 = a2
3a1 + b1 + 5 = b2
575
10.4 Nonhomogeneous Linear Systems
from which we obtain a3 = −1/4, b3 = 3/4, a2 = 1/4, b2 = −1/4, a1 = −2, and b1 = 3/4. Then
X(t) = c1
(
1
−1
)
e−2t + c2
(
1
1
)
e4t +
(− 14
3
4
)
t2 +
( 1
4
− 14
)
t +
(−2
3
4
)
.
4. Solving
det(A− λI) =
∣∣∣∣ 1− λ −44 1− λ
∣∣∣∣ = λ2 − 2λ + 17 = 0
we obtain eigenvalues λ1 = 1 + 4i and λ2 = 1− 4i. Corresponding eigenvectors are
K1 =
(
i
1
)
and K2 =
(−i
1
)
.
Thus
Xc = c1
[(
0
1
)
cos 4t +
(−1
0
)
sin 4t
]
et + c2
[(−1
0
)
cos 4t−
(
0
1
)
sin 4t
]
et
= c1
(− sin 4t
cos 4t
)
et + c2
(− cos 4t
− sin 4t
)
et.
Substituting
Xp =
(
a3
b3
)
t +
(
a2
b2
)
+
(
a1
b1
)
e6t
into the system yields
a3 − 4b3 = −4
4a3 + b3 = 1
a2 − 4b2 = a3
4a2 + b2 = b3
−5a1 − 4b1 = −9
4a1 − 5b1 = −1
from which we obtain a3 = 0, b3 = 1, a2 = 4/17, b2 = 1/17, a1 = 1, and b1 = 1. Then
X(t) = c1
(− sin 4t
cos 4t
)
et + c2
(− cos 4t
− sin 4t
)
et +
(
0
1
)
t +
( 4
17
1
17
)
+
(
1
1
)
e6t.
5. Solving
det(A− λI) =
∣∣∣∣ 4− λ 139 6− λ
∣∣∣∣ = λ2 − 10λ + 21 = (λ− 3)(λ− 7) = 0
we obtain the eigenvalues λ1 = 3 and λ2 = 7. Corresponding eigenvectors are
K1 =
(
1
−3
)
and K2 =
(
1
9
)
.
Thus
Xc = c1
(
1
−3
)
e3t + c2
(
1
9
)
e7t.
Substituting
Xp =
(
a1
b1
)
et
into the system yields
3a1 +
1
3
b1 = 3
9a1 + 5b1 = −10
from which we obtain a1 = 55/36 and b1 = −19/4. Then
X(t) = c1
(
1
−3
)
e3t + c2
(
1
9
)
e7t +
( 55
36
− 194
)
et.
576
10.4 Nonhomogeneous Linear Systems
6. Solving
det(A− λI) =
∣∣∣∣−1− λ 5−1 1− λ
∣∣∣∣ = λ2 + 4 = 0
we obtain the eigenvalues λ1 = 2i and λ2 = −2i. Corresponding eigenvectors are
K1 =
(
5
1 + 2i
)
and K2 =
(
5
1− 2i
)
.
Thus
Xc = c1
(
5 cos 2t
cos 2t− 2 sin 2t
)
+ c2
(
5 sin 2t
2 cos 2t + sin 2t
)
.
Substituting
Xp =
(
a2
b2
)
cos t +
(
a1
b1
)
sin t
into the system yields
−a2 + 5b2 − a1 = 0
−a2 + b2 − b1 − 2 = 0
−a1 + 5b1 + a2 + 1 = 0
−a1 + b1 + b2 = 0
from which we obtain a2 = −3, b2 = −2/3, a1 = −1/3, and b1 = 1/3. Then
X(t) = c1
(
5 cos 2t
cos 2t− 2 sin 2t
)
+ c2
(
5 sin 2t
2 cos 2t + sin 2t
)
+
( −3
− 23
)
cos t +
(− 13
1
3
)
sin t.
7. Solving
det(A− λI) =
∣∣∣∣∣∣∣
1− λ 1 1
0 2− λ 3
0 0 5− λ
∣∣∣∣∣∣∣ = (1− λ)(2− λ)(5− λ) = 0
we obtain the eigenvalues λ1 = 1, λ2 = 2, and λ3 = 5. Corresponding eigenvectors are
K1 =
 10
0
 , K2 =
 11
0
 and K3 =
 12
2
 .
Thus
Xc = C1
 10
0
 et + C2
 11
0
 e2t + C3
 12
2
 e5t.
Substituting
Xp =
 a1b1
c1
 e4t
into the system yields
−3a1 + b1 + c1 = −1
−2b1 + 3c1 = 1
c1 = −2
577
10.4 Nonhomogeneous Linear Systems
from which we obtain c1 = −2, b1 = −7/2, and a1 = −3/2. Then
X(t) = C1
 10
0
 et + C2
 11
0
 e2t + C3
 12
2
 e5t +
−
3
2
− 72
−2
 e4t.
8. Solving
det(A− λI) =
∣∣∣∣∣∣∣
−λ 0 5
0 5− λ 0
5 0 −λ
∣∣∣∣∣∣∣ = −(λ− 5)2(λ + 5) = 0
we obtain the eigenvalues λ1 = 5, λ2 = 5, and λ3 = −5. Corresponding eigenvectors are
K1 =
 10
0
 , K2 =
 11
1
 and K3 =
 10
−1
 .
Thus
Xc = C1
 10
1
 e5t + C2
 11
1
 e5t + C3
 10
−1
 e−5t.
Substituting
Xp =
 a1b1
c1

into the system yields
5c1 = −5
5b1 = 10
5a1 = −40
from which we obtain c1 = −1, b1 = 2, and a1 = −8. Then
X(t) = C1
 10
1
 e5t + C2
 11
1
 e5t + C3
 10
−1
 e−5t +
−82
−1
 .
9. Solving
det(A− λI) =
∣∣∣∣−1− λ −23 4− λ
∣∣∣∣ = λ2 − 3λ + 2 = (λ− 1)(λ− 2) = 0
we obtain the eigenvalues λ1 = 1 and λ2 = 2. Corresponding eigenvectors are
K1 =
(
1
−1
)
and K2 =
(−4
6
)
.
Thus
Xc = c1
(
1
−1
)
et + c2
(−4
6
)
e2t.
Substituting
Xp =
(
a1
b1
)
578
10.4 Nonhomogeneous Linear Systems
into the system yields
−a1 − 2b1 = −3
3a1 + 4b1 = −3
from which we obtain a1 = −9 and b1 = 6. Then
X(t) = c1
(
1
−1
)
et + c2
(−4
6
)
e2t +
(−9
6
)
.
Setting
X(0) =
(−4
5
)
we obtain
c1 − 4c2 − 9 = −4
−c1 + 6c2 + 6 = 5.
Then c1 = 13 and c2 = 2 so
X(t) = 13
(
1
−1
)
et + 2
(−4
6
)
e2t +
(−9
6
)
.
10. (a) Let I =
(
i2
i3
)
so that
I′ =
(−2 −2
−2 −5
)
I +
(
60
60
)
and
Ic = c1
(
2
−1
)
e−t + c2
(
1
2
)
e−6t.
If Ip =
(
a1
b1
)
then Ip =
(
30
0
)
so that
I = c1
(
2
−1
)
e−t + c2
(
1
2
)
e−6t +
(
30
0
)
.
For I(0) =
(
0
0
)
we find c1 = −12 and c2 = −6.
(b) i1(t) = i2(t) + i3(t) = −12e−t − 18e−6t + 30.
11. From
X′ =
(
3 −3
2 −2
)
X +
(
4
−1
)
we obtain
Xc = c1
(
1
1
)
+ c2
(
3
2
)
et.
Then
Φ =
(
1 3et
1 2et
)
and Φ−1 =
( −2 3
e−t −e−t
)
so that
U =
∫
Φ−1F dt =
∫ ( −11
5e−t
)
dt =
( −11t
−5e−t
)
and
Xp = ΦU =
(−11
−11
)
t +
(−15
−10
)
.
579
10.4 Nonhomogeneous Linear Systems
12. From
X′ =
(
2 −1
3 −2
)
X +
(
0
4
)
t
we obtain
Xc = c1
(
1
1
)
et + c2
(
1
3
)
e−t.
Then
Φ =
(
et e−t
et 3e−t
)
and Φ−1 =
( 3
2e
−t − 12e−t
− 12et 12et
)
so that
U =
∫
Φ−1F dt =
∫ (−2te−t
2tet
)
dt =
(
2te−t + 2e−t
2tet − 2et
)
and
Xp = ΦU =
(
4
8
)
t +
(
0
−4
)
.
13. From
X′ =
(
3 −5
3
4 −1
)
X +
(
1
−1
)
et/2
we obtain
Xc = c1
(
10
3
)
e3t/2 + c2
(
2
1
)
et/2.
Then
Φ =
(
10e3t/2 2et/2
3e3t/2 et/2
)and Φ−1 =
( 1
4e
−3t/2 − 12e−3t/2
− 34e−t/2 52e−t/2
)
so that
U =
∫
Φ−1F dt =
∫ ( 3
4e
−t
− 134
)
dt =
(− 34e−t
− 134 t
)
and
Xp = ΦU =
(− 132
− 134
)
tet/2 +
(− 152
− 94
)
et/2.
14. From
X′ =
(
2 −1
4 2
)
X +
(
sin 2t
2 cos 2t
)
we obtain
Xc = c1
(− sin 2t
2 cos 2t
)
e2t + c2
(
cos 2t
2 sin 2t
)
e2t.
Then
Φ =
(−e2t sin 2t e2t cos 2t
2e2t cos 2t 2e2t sin 2t
)
and Φ−1 =
(− 12e−2t sin 2t 14e−2t cos 2t
1
2e
−2t cos 2t 14e
−2t sin 2t
)
so that
U =
∫
Φ−1F dt =
∫ ( 1
2 cos 4t
1
2 sin 4t
)
dt =
( 1
8 sin 4t
− 18 cos 4t
)
and
Xp = ΦU =
(− 18 sin 2t cos 4t− 18 cos 2t cos 4t
1
4 cos 2t sin 4t− 14 sin 2t cos 4t
)
e2t.
580
10.4 Nonhomogeneous Linear Systems
15. From
X′ =
(
0 2
−1 3
)
X +
(
1
−1
)
et
we obtain
Xc = c1
(
2
1
)
et + c2
(
1
1
)
e2t.
Then
Φ =
(
2et e2t
et e2t
)
and Φ−1 =
(
e−t −e−t
−e−2t 2e−2t
)
so that
U =
∫
Φ−1F dt =
∫ (
2
−3e−t
)
dt =
(
2t
3e−t
)
and
Xp = ΦU =
(
4
2
)
tet +
(
3
3
)
et.
16. From
X′ =
(
0 2
−1 3
)
X +
(
2
e−3t
)
we obtain
Xc = c1
(
2
1
)
et + c2
(
1
1
)
e2t.
Then
Φ =
(
2et e2t
et e2t
)
and Φ−1 =
(
e−t −e−t
−e−2t 2e−2t
)
so that
U =
∫
Φ−1F dt =
∫ (
2e−t − e−4t
−2e−2t + 2e−5t
)
dt =
(−2e−t + 14e−4t
e−2t − 25e−5t
)
and
Xp = ΦU =
(
1
10e
−3t − 3
− 320e−3t − 1
)
.
17. From
X′ =
(
1 8
1 −1
)
X +
(
12
12
)
t
we obtain
Xc = c1
(
4
1
)
e3t + c2
(−2
1
)
e−3t.
Then
Φ =
(
4e3t −2e−3t
e3t e−3t
)
and Φ−1 =
(
1
6e
−3t 1
3e
−3t
− 16e3t 23e3t
)
so that
U =
∫
Φ−1F dt =
∫ (
6te−3t
6te3t
)
dt =
(−2te−3t − 23e−3t
2te3t − 23e3t
)
and
Xp = ΦU =
(−12
0
)
t +
(− 43
− 43
)
.
581
10.4 Nonhomogeneous Linear Systems
18. From
X′ =
(
1 8
1 −1
)
X +
(
e−t
tet
)
we obtain
Xc = c1
(
4
1
)
e3t + c2
(−2
1
)
e−3t.
Then
Φ =
(
4e3t −2e3t
e3t e−3t
)
and Φ−1 =
(
1
6e
−3t 1
3e
−3t
− 16e3t 23e3t
)
so that
U =
∫
Φ−1F dt =
∫ ( 1
6e
−4t + 13 te
−2t
− 16e2t + 23 te4t
)
dt =
(− 124e−4t − 16 te−2t − 112e−2t
− 112e2t + 16 te4t − 124e4t
)
and
Xp = ΦU =
( −tet − 14et
− 18e−t − 18et
)
.
19. From
X′ =
(
3 2
−2 −1
)
X +
(
2
1
)
e−t
we obtain
Xc = c1
(
1
−1
)
et + c2
[(
1
−1
)
tet +
(
0
1
2
)
et
]
.
Then
Φ =
(
et tet
−et 12et − tet
)
and Φ−1 =
(
e−t − 2te−t −2te−t
2e−t 2e−t
)
so that
U =
∫
Φ−1F dt =
∫ (
2e−2t − 6te−2t
6e−2t
)
dt =
( 1
2e
−2t + 3te−2t
−3e−2t
)
and
Xp = ΦU =
( 1
2
−2
)
e−t.
20. From
X′ =
(
3 2
−2 −1
)
X +
(
1
1
)
we obtain
Xc = c1
(
1
−1
)
et + c2
[(
1
−1
)
tet +
(
0
1
2
)
et
]
.
Then
Φ =
(
et tet
−et 12et − tet
)
and Φ−1 =
(
e−t − 2te−t −2te−t
2e−t 2e−t
)
so that
U =
∫
Φ−1F dt =
∫ (
e−t − 4te−t
2e−t
)
dt =
(
3e−t + 4te−t
−2e−t
)
and
Xp = ΦU =
(
3
−5
)
.
582
10.4 Nonhomogeneous Linear Systems
21. From
X′ =
(
0 −1
1 0
)
X +
(
sec t
0
)
we obtain
Xc = c1
(
cos t
sin t
)
+ c2
(
sin t
− cos t
)
.
Then
Φ =
(
cos t sin t
sin t − cos t
)
and Φ−1 =
(
cos t sin t
sin t − cos t
)
so that
U =
∫
Φ−1F dt =
∫ (
1
tan t
)
dt =
(
t
− ln | cos t|
)
and
Xp = ΦU =
(
t cos t− sin t ln | cos t|
t sin t + cos t ln | cos t|
)
.
22. From
X′ =
(
1 −1
1 1
)
X +
(
3
3
)
et
we obtain
Xc = c1
(− sin t
cos t
)
et + c2
(
cos t
sin t
)
et.
Then
Φ =
(− sin t cos t
cos t sin t
)
et and Φ−1 =
(− sin t cos t
cos t sin t
)
e−t
so that
U =
∫
Φ−1F dt =
∫ (−3 sin t + 3 cos t
3 cos t + 3 sin t
)
dt =
(
3 cos t + 3 sin t
3 sin t− 3 cos t
)
and
Xp = ΦU =
(−3
3
)
et.
23. From
X′ =
(
1 −1
1 1
)
X +
(
cos t
sin t
)
et
we obtain
Xc = c1
(− sin t
cos t
)
et + c2
(
cos t
sin t
)
et.
Then
Φ =
(− sin t cos t
cos t sin t
)
et and Φ−1 =
(− sin t cos t
cos t sin t
)
e−t
so that
U =
∫
Φ−1F dt =
∫ (
0
1
)
dt =
(
0
t
)
and
Xp = ΦU =
(
cos t
sin t
)
tet.
583
10.4 Nonhomogeneous Linear Systems
24. From
X′ =
(
2 −2
8 −6
)
X +
(
1
3
)
1
t
e−2t
we obtain
Xc = c1
(
1
2
)
e−2t + c2
[(
1
2
)
te−2t +
( 1
2
1
2
)
e−2t
]
.
Then
Φ =
(
1 t + 12
2 2t + 12
)
e−2t and Φ−1 =
(−4t− 1 2t + 1
4 −2
)
e2t
so that
U =
∫
Φ−1F dt =
∫ (
2 + 2/t
−2/t
)
dt =
(
2t + 2 ln t
−2 ln t
)
and
Xp = ΦU =
(
2t + ln t− 2t ln t
4t + 3 ln t− 4t ln t
)
e−2t.
25. From
X′ =
(
0 1
−1 0
)
X +
(
0
sec t tan t
)
we obtain
Xc = c1
(
cos t
− sin t
)
+ c2
(
sin t
cos t
)
.
Then
Φ =
(
cos t sin t
− sin t cos t
)
t and Φ−1 =
(
cos t − sin t
sin t cos t
)
so that
U =
∫
Φ−1F dt =
∫ (− tan2 t
tan t
)
dt =
(
t− tan t
− ln | cos t|
)
and
Xp = ΦU =
(
cos t
− sin t
)
t +
( − sin t
sin t tan t
)
−
(
sin t
cos t
)
ln | cos t|.
26. From
X′ =
(
0 1
−1 0
)
X +
(
1
cot t
)
we obtain
Xc = c1
(
cos t
− sin t
)
+ c2
(
sin t
cos t
)
.
Then
Φ =
(
cos t sin t
− sin t cos t
)
and Φ−1 =
(
cos t − sin t
sin t cos t
)
so that
U =
∫
Φ−1F dt =
∫ (
0
csc t
)
dt =
(
0
ln | csc t− cot t|
)
and
Xp = ΦU =
(
sin t ln | csc t− cot t|
cos t ln | csc t− cot t|
)
.
584
10.4 Nonhomogeneous Linear Systems
27. From
X′ =
(
1 2
− 12 1
)
X +
(
csc t
sec t
)
et
we obtain
Xc = c1
(
2 sin t
cos t
)
et + c2
(
2 cos t
− sin t
)
et.
Then
Φ =
(
2 sin t 2 cos t
cos t − sin t
)
et and Φ−1 =
(
1
2 sin t cos t
1
2 cos t − sin t
)
e−t
so that
U =
∫
Φ−1F dt =
∫ ( 3
2
1
2 cot t− tan t
)
dt =
(
3
2 t
1
2 ln | sin t|+ ln | cos t|
)
and
Xp = ΦU =
(
3 sin t
3
2 cos t
)
tet +
(
cos t
− 12 sin t
)
et ln | sin t|+
(
2 cos t
− sin t
)
et ln | cos t|.
28. From
X′ =
(
1 −2
1 −1
)
X +
(
tan t
1
)
we obtain
Xc = c1
(
cos t− sin t
cos t
)
+ c2
(
cos t + sin t
sin t
)
.
Then
Φ =
(
cos t− sin t cos t + sin t
cos t sin t
)
and Φ−1 =
(− sin t cos t + sin t
cos t sin t− cos t
)
so that
U =
∫
Φ−1F dt =
∫ (
2 cos t + sin t− sec t
2 sin t− cos t
)
dt =
(
2 sin t− cos t− ln | sec t + tan t|
−2 cos t− sin t
)
and
Xp = ΦU =
(
3 sin t cos t− cos2 t− 2 sin2 t + (sin t− cos t) ln | sec t + tan t|
sin2 t− cos2 t− cos t(ln | sec t + tan t|)
)
.
29. From
X′ =
 1 1 01 1 0
0 0 3
X +
 e
t
e2t
te3t

we obtain
Xc = c1
 1−1
0
 + c2
 11
0
 e2t + c3
 00
1
 e3t.
Then
Φ =
 1 e
2t 0
−1 e2t 0
0 0 e3t
 and Φ−1 =

1
2 − 12 0
1
2e
−2t 1
2e
−2t 0
0 0 e−3t

so that
U =
∫
Φ−1F dt =
∫ 
1
2e
t − 12e2t
1
2e
−t + 12
t
 dt =

1
2e
t − 14e2t
− 12e−t + 12 t
1
2 t
2

585
10.4 Nonhomogeneous Linear Systems
and
Xp = ΦU =

− 14e2t + 12 te2t
−et + 14e2t + 12 te2t
1
2 t
2e3t
 .
30. From
X′ =
 3 −1 −11 1 −1
1 −1 1
X +
 0t
2et
we obtain
Xc = c1
 11
1
 et + c2
 11
0
 e2t + c3
 10
1
 e2t.
Then
Φ =
 e
t e2t e2t
et e2t 0
et 0 e2t
 and Φ−1 =
−e
−t e−t e−t
e−2t 0 −e−2t
e−2t −e−2t 0

so that
U =
∫
Φ−1F dt =
∫  te
−t + 2
−2e−t
−te−2t
 dt =
−te
−t − e−t + 2t
2e−t
1
2 te
−2t + 14e
−2t

and
Xp = ΦU =
−
1
2
−1
− 12
 t +
−
3
4
−1
− 34
 +
 22
0
 et +
 22
2
 tet.
31. From
X′ =
(
3 −1
−1 3
)
X +
(
4e2t
4e4t
)
we obtain
Φ =
(−e4t e2t
e4t e2t
)
, Φ−1 =
(− 12e−4t 12e−4t
1
2e
−2t 1
2e
−2t
)
,
and
X = ΦΦ−1(0)X(0) + Φ
∫ t
0
Φ−1F ds = Φ ·
(
0
1
)
+ Φ ·
(
e−2t + 2t− 1
e2t + 2t− 1
)
=
(
2
2
)
te2t +
(−1
1
)
e2t +
(−2
2
)
te4t +
(
2
0
)
e4t.
32. From
X′ =
(
1 −1
1 −1
)
X +
(
1/t
1/t
)
we obtain
Φ =
(
1 1 + t
1 t
)
, Φ−1 =
(−t 1 + t
1 −1
)
,
and
X = ΦΦ−1(1)X(1) + Φ
∫ t
1
Φ−1F ds = Φ ·
(−4
3
)
+ Φ ·
(
ln t
0
)
=
(
3
3
)
t−
(
1
4
)
+
(
1
1
)
ln t.
586
10.4 Nonhomogeneous Linear Systems
33. Let I =
(
i1
i2
)
so that
I′ =
(−11 3
3 −3
)
I +
(
100 sin t
0
)
and
Ic = c1
(
1
3
)
e−2t + c2
(
3
−1
)
e−12t.
Then
Φ =
(
e−2t 3e−12t
3e−2t −e−12t
)
, Φ−1 =
(
1
10e
2t 3
10e
2t
3
10e
12t − 110e12t
)
,
U =
∫
Φ−1F dt =
∫ (
10e2t sin t
30e12t sin t
)
dt =
(
2e2t(2 sin t− cos t)
6
29e
12t(12 sin t− cos t)
)
,
and
Ip = ΦU =
(
332
29 sin t− 7629 cos t
276
29 sin t− 16829 cos t
)
so that
I = c1
(
1
3
)
e−2t + c2
(
3
−1
)
e−12t + Ip.
If I(0) =
(
0
0
)
then c1 = 2 and c2 = 629 .
34. (a) The eigenvalues are 0, 1, 3, and 4, with corresponding eigenvectors
−6
−4
1
2
 ,

2
1
0
0
 ,

3
1
2
1
 , and

−1
1
0
0
 .
(b) Φ =

−6 2et 3e3t −e4t
−4 et e3t e4t
1 0 2e3t 0
2 0 e3t 0
, Φ−1 =

0 0 − 13 23
1
3 e
−t 1
3 e
−t −2e−t 83 e−t
0 0 23 e
−3t − 13 e−3t
− 13 e−4t 23 e−4t 0 13 e−4t

(c) Φ−1(t)F(t) =

2
3 − 13 e2t
1
3 e
−2t + 83 e
−t − 2et + 13 t
− 13 e−3t + 23e−t
2
3 e
−5t + 13 e
−4t − 13 te−3t
,
∫
Φ−1(t)F(t)dt =

− 16 e2t + 23 t
− 16 e−2t − 83 e−t − 2et + 16 t2
1
9 e
−3t − 23 e−t
− 215 e−5t − 112 e−4t + 127 e−3t + 19 te−3t
,
Xp(t) = Φ(t)
∫
Φ−1(t)F(t)dt =

−5e2t − 15 e−t − 127 et − 19 tet + 13 t2et − 4t− 5912
−2e2t − 310 e−t + 127et + 19 tet + 16 t2et − 83 t− 9536
− 32 e2t + 23 t + 29
−e2t + 43 t− 19
,
587
10.4 Nonhomogeneous Linear Systems
Xc(t) = Φ(t)C =

−6c1 + 2c2et + 3c3e3t − c4e4t
−4c1 + c2et + c3e3t + c4e4t
c1 + 2c3e3t
2c1 + c3e3t
,
X(t) = Φ(t)C + Φ(t)
∫
Φ−1(t)F(t)dt
=

−6c1 + 2c2et + 3c3e3t − c4e4t
−4c1 + c2et + c3e3t + c4e4t
c1 + 2c3e3t
2c1 + c3e3t
 +

−5e2t − 15 e−t − 127 et − 19 tet + 13 t2et − 4t− 5912
−2e2t − 310 e−t + 127et + 19 tet + 16 t2et − 83 t− 9536
− 32 e2t + 23 t + 29
−e2t + 43 t− 19

(d) X(t) = c1

−6
−4
1
2
 + c2

2
1
0
0
 et + c3

3
1
2
1
 e3t + c4

−1
1
0
0
 e4t
+

−5e2t − 15 e−t − 127 et − 19 tet + 13 t2et − 4t− 5912
−2e2t − 310 e−t + 127et + 19 tet + 16 t2et − 83 t− 9536
− 32 e2t + 23 t + 29
−e2t + 43 t− 19

35. λ1 = −1, λ2 = −2, K1 =
(
1
3
)
, K2 =
(
2
7
)
, P =
(
1 2
3 7
)
, P−1 =
(
7 −2
−3 1
)
, P−1F =
(
34
−14
)
;
Y′ =
(−1 0
0 −2
)
Y +
(
34
−14
)
y1 = 34 + c1e−t, y2 = −7 + c2e−2t
X = PY =
(
1 2
3 7
) (
34 + c1e−t
−7 + c2e−2t
)
=
(
20 + c1e−t + 2c2e−2t
53 + 3c1e−t + 7c2e−2t
)
= c1
(
1
3
)
e−t + c2
(
2
7
)
e−2t +
(
20
53
)
36. λ1 = −1, λ2 = 4, K1 =
(
3
−2
)
, K2 =
(
1
1
)
, P =
(
3 1
−2 1
)
, P−1 =
1
5
(
1 −1
2 3
)
, P−1F =
(
0
et
)
;
Y′ =
(−1 0
0 4
)
Y +
(
0
et
)
y1 = c1e−t, y2 = −13e
t + c2e4t
X = PY =
(
3 1
−2 1
) (
c1e
−t
− 13et + c2e4t
)
=
(− 13et + 3c1e−t + c2e4t
− 13et − 2c1e−t + c2e4t
)
= c1
(
3
−2
)
e−t + c2
(
1
1
)
e4t − 1
3
(
1
1
)
et
37. λ1 = 0, λ2 = 10, K1 =
(
1
−1
)
, K2 =
(
1
1
)
, P =
(
1 1
−1 1
)
, P−1 =
1
2
(
1 −1
1 1
)
, P−1F =
(
t− 4
t + 4
)
;
Y′ =
(
0 0
0 10
)
Y +
(
t− 4
t + 4
)
y1 =
1
2
t2 − 4t + c1, y2 = − 110 t−
41
100
+ c2e10t
588
10.5 Matrix Exponential
X = PY =
(
1 1
−1 1
) ( 1
2 t
2 − 4t + c1
− 110 t− 41100 + c2e10t
)
=
( 1
2 t
2 − 4110 t− 41100 + c1 + c2e10t
− 12 t2 + 3910 t− 41100 − c1 + c2e10t
)
= c1
(
1
−1
)
+ c2
(
1
1
)
e10t +
1
2
(
1
−1
)
t2 +
1
10
(−41
39
)
t− 41
100
(
1
1
)
38. λ1 = −1, λ2 = 1, K1 =
(
1
−1
)
, K2 =
(
1
1
)
, P =
(
1 1
−1 1
)
, P−1 =
1
2
(
1 −1
1 1
)
, P−1F =
(
2− 4e−2t
2 + 4e−2t
)
;
Y′ =
(−1 0
0 1
)
Y +
(
2− 4e−2t
2 + 4e−2t
)
y1 = 2 + 4e−2t + c1e−t, y2 = −2− 43e
−2t + c2et
X = PY =
(
1 1
−1 1
) (
2 + 4e−2t + c1e−t
−2− 43e−2t + c2et
)
=
( 8
3e
−2t + c1e−t + c2et
−4− 163 e−2t − c1e−t + c2et
)
= c1
(
1
−1
)
e−t + c2
(
1
1
)
et +
8
3
(
1
−2
)
e−2t +
(
0
−4
)
EXERCISES 10.5
Matrix Exponential
1. For A =
(
1 0
0 2
)
we have
A2 =
(
1 0
0 2
) (
1 0
0 2
)
=
(
1 0
0 4
)
,
A3 = AA2 =
(
1 0
0 2
) (
1 0
0 4
)
=
(
1 0
0 8
)
,
A4 = AA3 =
(
1 0
0 2
) (
1 0
0 8
)
=
(
1 0
0 16
)
,
and so on. In general
Ak =
(
1 0
0 2k
)
for k = 1, 2, 3, . . . .
Thus
eAt = I +
A
1!
t +
A2
2!
t2 +
A3
3!
t3 + · · ·
=
(
1 0
0 1
)
+
1
1!
(
1 0
0 2
)
t +
1
2!
(
1 0
0 4
)
t2 +
1
3!
(
1 0
0 8
)
t3 + · · ·
=
 1 + t +
t2
2!
+
t3
3!
+ · · · 0
0 1 + t +
(2t)2
2!
+
(2t)3
3!
+ · · ·
 = ( et 0
0 e2t
)
589
10.5 Matrix Exponential
and
e−At =
(
e−t 0
0 e−2t
)
.
2. For A =
(
0 1
1 0
)
we have
A2 =
(
0 1
1 0
) (
0 1
1 0
)
=
(
1 0
0 1
)
= I
A3 = AA2 =
(
0 1
1 0
)
I =
(
0 1
1 0
)
= A
A4 = (A2)2 = I
A5 = AA4 = AI = A,
and so on. In general,
Ak =
{
A, k = 1, 3, 5, . . .
I, k = 2, 4, 6, . . . .
Thus
eAt = I +
A
1!
t +
A2
2!
t2 +
A3
3!
t3 + · · ·
= I + At +
1
2!
It2 +
1
3!
At3 + · · ·
= I
(
1 +
1
2!
t2 +
1
4!
t4 + · · ·
)
+ A
(
t +
1
3!
t3 +
1
5!
t5 + · · ·
)
= I cosh t + A sinh t =
(
cosh t sinh t
sinh t cosh t
)
and
e−At =
(
cosh(−t) sinh(−t)
sinh(−t) cosh(−t)
)
=
(
cosh t − sinh t
− sinh t cosh t
)
.
3. For
A =
 1 1 11 1 1
−2 −2 −2

we have
A2 =
 1 1 11 1 1
−2 −2 −2

 1 1 11 1 1
−2 −2 −2
 =
 0 0 00 0 0
0 0 0
 .
Thus, A3 = A4 = A5 = · · · = 0 and
eAt = I + At =
 1 0 00 1 0
0 0 1
 +
 t t tt t t
−2t −2t −2t
 =
 t + 1 t tt t + 1 t
−2t −2t −2t + 1
 .
590
10.5 Matrix Exponential
4. For
A =
 0 0 03 0 0
5 1 0

we have
A2 =
 0 0 03 0 0
5 1 0

 0 0 03 0 0
5 1 0
 =
 0 0 00 0 0
3 0 0

A3 = AA2 =
 0 0 03 0 0
5 1 0

 0 0 00 0 0
3 0 0
 =
 0 0 00 00
0 0 0
 .
Thus, A4 = A5 = A6 = · · · = 0 and
eAt = I + At +
1
2
A2t2
=
 1 0 00 1 0
0 0 1
 +
 0 0 03t 0 0
5t t 0
 +
 0 0 00 0 0
3
2 t
2 0 0
 =
 1 0 03t 1 0
3
2 t
2 + 5t t 1
 .
5. Using the result of Problem 1,
X =
(
et 0
0 e2t
) (
c1
c2
)
= c1
(
et
0
)
+ c2
(
0
et
)
.
6. Using the result of Problem 2,
X =
(
cosh t sinh t
sinh t cosh t
) (
c1
c2
)
= c1
(
cosh t
sinh t
)
+ c2
(
sinh t
cosh t
)
.
7. Using the result of Problem 3,
X =
 t + 1 t tt t + 1 t
−2t −2t −2t + 1

 c1c2
c3
 = c1
 t + 1t
−2t
 + c2
 tt + 1
−2t
 + c3
 tt
−2t + 1
 .
8. Using the result of Problem 4,
X =
 1 0 03t 1 0
3
2 t
2 + 5t t 1

 c1c2
c3
 = c1
 13t
3
2 t
2 + 5t
 + c2
 01
t
 + c3
 00
1
 .
9. To solve
X′ =
(
1 0
0 2
)
X +
(
3
−1
)
591
10.5 Matrix Exponential
we identify t0 = 0, F(t) =
(
3
−1
)
, and use the results of Problem 1 and equation (6) in the text.
X(t) = eAtC + eAt
∫ t
t0
e−AsF(s) ds
=
(
et 0
0 e2t
) (
c1
c2
)
+
(
et 0
0 e2t
) ∫ t
0
(
e−s 0
0 e−2s
) (
3
−1
)
ds
=
(
c1e
t
c2e
2t
)
+
(
et 0
0 e2t
) ∫ t
0
(
3e−s
−e−2s
)
ds
=
(
c1e
t
c2e
2t
)
+
(
et 0
0 e2t
) (−3e−s
1
2e
−2s
) ∣∣∣∣t
0
=
(
c1e
t
c2e
2t
)
+
(
et 0
0 e2t
) (−3e−t + 3
1
2e
−2t − 12
)
=
(
c1e
t
c2e
2t
)
+
(−3 + 3et
1
2 − 12e2t
)
= c3
(
1
0
)
et + c4
(
0
1
)
e2t +
( −3
1
2
)
.
10. To solve
X′ =
(
1 0
0 2
)
X +
(
t
e4t
)
we identify t0 = 0, F(t) =
(
t
e4t
)
, and use the results of Problem 1 and equation (6) in the text.
X(t) = eAtC + eAt
∫ t
t0
e−AsF(s) ds
=
(
et 0
0 e2t
) (
c1
c2
)
+
(
et 0
0 e2t
) ∫ t
0
(
e−s 0
0 e−2s
) (
s
e4s
)
ds
=
(
c1e
t
c2e
2t
)
+
(
et 0
0 e2t
) ∫ t
0
(
se−s
e2s
)
ds
=
(
c1e
t
c2e
2t
)
+
(
et 0
0 e2t
) (−se−s − e−s
1
2e
2s
) ∣∣∣∣t
0
=
(
c1e
t
c2e
2t
)
+
(
et 0
0 e2t
) (−te−t − e−t + 1
1
2e
2t − 12
)
=
(
c1e
t
c2e
2t
)
+
(−t− 1 + et
1
2e
4t − 12e2t
)
= c3
(
1
0
)
et + c4
(
0
1
)
e2t +
(−t− 1
1
2e
4t
)
.
11. To solve
X′ =
(
0 1
1 0
)
X +
(
1
1
)
592
10.5 Matrix Exponential
we identify t0 = 0, F(t) =
(
1
1
)
, and use the results of Problem 2 and equation (6) in the text.
X(t) = eAtC + eAt
∫ t
t0
e−AsF(s) ds
=
(
cosh t sinh t
sinh t cosh t
) (
c1
c2
)
+
(
cosh t sinh t
sinh t cosh t
) ∫ t
0
(
cosh s − sinh s
− sinh s cosh s
) (
1
1
)
ds
=
(
c1 cosh t + c2 sinh t
c1 sinh t + c2 cosh t
)
+
(
cosh t sinh t
sinh t cosh t
) ∫ t
0
(
cosh s− sinh s
− sinh s + cosh s
)
ds
=
(
c1 cosh t + c2 sinh t
c1 sinh t + c2 cosh t
)
+
(
cosh t sinh t
sinh t cosh t
) (
sinh s− cosh s
− cosh s + sinh s
) ∣∣∣∣t
0
=
(
c1 cosh t + c2 sinh t
c1 sinh t + c2 cosh t
)
+
(
cosh t sinh t
sinh t cosh t
) (
sinh t− cosh t + 1
− cosh t + sinh t + 1
)
=
(
c1 cosh t + c2 sinh t
c1 sinh t + c2 cosh t
)
+
(
sinh2 t− cosh2 t + cosh t + sinh t
sinh2 t− cosh2 t + sinh t + cosh t
)
= c1
(
cosh t
sinh t
)
+ c2
(
sinh t
cosh t
)
+
(
cosh t
sinh t
)
+
(
sinh t
cosh t
)
−
(
1
1
)
= c3
(
cosh t
sinh t
)
+ c4
(
sinh t
cosh t
)
−
(
1
1
)
.
12. To solve
X′ =
(
0 1
1 0
)
X +
(
cosh t
sinh t
)
we identify t0 = 0, F(t) =
(
cosh t
sinh t
)
, and use the results of Problem 2 and equation (6) in the text.
X(t) = eAtC + eAt
∫ t
t0
e−AsF(s) ds
=
(
cosh t sinh t
sinh t cosh t
) (
c1
c2
)
+
(
cosh t sinh t
sinh t cosh t
) ∫ t
0
(
cosh s − sinh s
− sinh s cosh s
) (
cosh s
sinh s
)
ds
=
(
c1 cosh t + c2 sinh t
c1 sinh t + c2 cosh t
)
+
(
cosh t sinh t
sinh t cosh t
) ∫ t
0
(
1
0
)
ds
=
(
c1 cosh t + c2 sinh t
c1 sinh t + c2 cosh t
)
+
(
cosh t sinh t
sinh t cosh t
) (
s
0
) ∣∣∣∣t
0
=
(
c1 cosh t + c2 sinh t
c1 sinh t + c2 cosh t
)
+
(
cosh t sinh t
sinh t cosh t
) (
t
0
)
=
(
c1 cosh t + c2 sinh t
c1 sinh t + c2 cosh t
)
+
(
t cosh t
t sinh t
)
= c1
(
cosh t
sinh t
)
+ c2
(
sinh t
cosh t
)
+ t
(
cosh t
sinh t
)
.
13. We have
X(0) = c1
 10
0
 + c2
 01
0
 + c3
 00
1
 =
 c1c2
c3
 =
 1−4
6
 .
593
10.5 Matrix Exponential
Thus, the solution of the initial-value problem is
X =
 t + 1t
−2t
− 4
 tt + 1
−2t
 + 6
 tt
−2t + 1
 .
14. We have
X(0) = c3
(
1
0
)
+ c4
(
0
1
)
+
(−3
1
2
)
=
(
c3 − 3
c4 + 12
)
=
(
4
3
)
.
Thus, c3 = 7 and c4 = 52 , so
X = 7
(
1
0
)
et +
5
2
(
0
1
)
e2t +
(−3
1
2
)
.
15. From sI−A =
(
s− 4 −3
4 s + 4
)
we find
(sI−A)−1 =

3/2
s− 2 −
1/2
s + 2
3/4
s− 2 −
3/4
s + 2
−1
s− 2 +
1
s + 2
−1/2
s− 2 +
3/2
s + 2

and
eAt =
(
3
2e
2t − 12e−2t 34e2t − 34e−2t
−e2t + e−2t − 12e2t + 32e−2t
)
.
The general solution of the system is then
X = eAtC =
(
3
2e
2t − 12e−2t 34e2t − 34e−2t
−e2t + e−2t − 12e2t + 32e−2t
)(
c1
c2
)
= c1
( 3
2
−1
)
e2t + c1
(− 12
1
)
e−2t + c2
( 3
4
− 12
)
e2t + c2
(− 34
3
2
)
e−2t
=
(1
2
c1 +
1
4
c2
) ( 3
−2
)
e2t +
(
−1
2
c1 − 34c2
) ( 1
−2
)
e−2t
= c3
(
3
−2
)
e2t + c4
(
1
−2
)
e−2t.
16. From sI−A =
(
s− 4 2
−1 s− 1
)
we find
(sI−A)−1 =

2
s− 3 −
1
s− 2 −
2
s− 3 +
2
s− 2
1
s− 3 −
1
s− 2
−1
s− 3 +
2
s− 2

and
eAt =
(
2e3t − e2t −2e3t + 2e2t
e3t − e2t −e3t + 2e2t
)
.
The general solution of the system is then
594
10.5 Matrix Exponential
X = eAtC =
(
2e3t − e2t −2e3t + 2e2t
e3t − e2t −e3t + 2e2t
) (
c1
c2
)
= c1
(
2
1
)
e3t + c1
(−1
−1
)
e2t + c2
(−2
−1
)
e3t + c2
(
2
2
)
e2t
= (c1 − c2)
(
2
1
)
e3t + (−c1 + 2c2)
(
1
1
)
e2t
= c3
(
2
1
)
e3t + c4
(
1
1
)
e2t.
17. From sI−A =
(
s− 5 9
−1 s + 1
)
we find
(sI−A)−1 =

1
s− 2 +
3
(s− 2)2 −
9
(s− 2)2
1
(s− 2)2
1
s− 2 −
3
(s− 2)2

and
eAt =
(
e2t + 3te2t −9te2t
te2t e2t − 3te2t
)
.
The general solution of the system is then
X = eAtC =
(
e2t + 3te2t −9te2t
te2t e2t − 3te2t
) (
c1
c2
)
= c1
(
1
0
)
e2t + c1
(
3
1
)
te2t + c2
(
0
1
)
e2t + c2
(−9
−3
)
te2t
= c1
(
1 + 3t
t
)
e2t + c2
( −9t
1− 3t
)
e2t.
18. From sI−A =
(
s −1
2 s + 2
)
we find
(sI−A)−1 =

s + 1 + 1
(s + 1)2 + 1
1
(s + 1)2 + 1
−2
(s + 1)2 + 1
s + 1− 1
(s + 1)2 + 1

and
eAt =
(
e−t cos t + e−t sin t e−t sin t
−2e−t sin t e−t cos t− e−t sin t
)
.
The general solution of the system is then
X = eAtC =
(
e−t cos t + e−t sin t e−t sin t
−2e−t sin t e−t cos t− e−t sin t
) (
c1
c2
)
= c1
(
1
0
)
e−t cos t + c1
(
1
−2
)
e−t sin t + c2
(
0
1
)
e−t cos t + c2
(
1
−1
)
e−t sin t
595
10.5 Matrix Exponential
= c1
(
cos t + sin t
−2 sin t
)
e−t + c2(
sin t
cos t− sin t
)
e−t.
19. The eigenvalues are λ1 = 1 and λ2 = 6. This leads to the system
et = b0 + b1
e6t = b0 + 6b1,
which has the solution b0 = 65e
t − 15e6t and b1 = − 15et + 15e6t. Then
eAt = b0I + b1A =
(
4
5e
t + 15e
6t 2
5e
t − 25e6t
2
5e
t − 25e6t 15et + 45e6t
)
.
The general solution of the system is then
X = eAtC =
(
4
5e
t + 15e
6t 2
5e
t − 25e6t
2
5e
t − 25e6t 15et + 45e6t
)(
c1
c2
)
= c1
( 4
5
2
5
)
et + c1
( 1
5
− 25
)
e6t + c2
( 2
5
1
5
)
et + c2
(− 25
4
5
)
e6t
=
(2
5
c1 +
1
5
c2
) ( 2
1
)
et +
(1
5
c1 − 25c2
) ( 1
−2
)
e6t
= c3
(
2
1
)
et + c4
(
1
−2
)
e6t.
20. The eigenvalues are λ1 = 2 and λ2 = 3. This leads to the system
e2t = b0 + 2b1
e3t = b0 + 3b1,
which has the solution b0 = 3e2t − 2e3t and b1 = −e2t + e3t. Then
eAt = b0I + b1A =
(
2e2t − e3t −2e2t + 2e3t
e2t − e3t −e2t + 2e3t
)
.
The general solution of the system is then
X = eAtC =
(
2e2t − e3t −2e2t + 2e3t
e2t − e3t −e2t + 2e3t
) (
c1
c2
)
= c1
(
2
1
)
e2t + c1
(−1
−1
)
e3t + c2
(−2
−1
)
e2t + c2
(
2
2
)
e3t
= (c1 − c2)
(
2
1
)
e2t + (−c1 + 2c2)
(
1
1
)
e3t
= c3
(
2
1
)
e2t + c4
(
1
1
)
e3t.
596
10.5 Matrix Exponential
21. The eigenvalues are λ1 = −1 and λ2 = 3. This leads to the system
e−t = b0 − b1
e3t = b0 + 3b1,
which has the solution b0 = 34e
−t + 14e
3t and b1 = − 14e−t + 14e3t. Then
eAt = b0I + b1A =
(
e3t −2e−t + 2e3t
0 e−t
)
.
The general solution of the system is then
X = eAtC =
(
e3t −2e−t + 2e3t
0 e−t
)(
c1
c2
)
= c1
(
1
0
)
e3t + c2
(−2
1
)
e−t + c2
(
2
0
)
e3t
= c2
(−2
1
)
e−t + (c1 + 2c2)
(
1
0
)
e3t
= c3
(−2
1
)
e−t + c4
(
1
0
)
e3t.
22. The eigenvalues are λ1 = 14 and λ2 =
1
2 . This leads to the system
et/4 = b0 +
1
4
b1
et/2 = b0 +
1
2
b1,
which has the solution b0 = 2et/4 + et/2 and b1 = −4et/4 + 4et/2. Then
eAt = b0I + b1A =
(−2et/4 + 3et/2 6et/4 − 6et/2
−et/4 + et/2 3et/4 − 2et/2
)
.
The general solution of the system is then
X = eAtC =
(−2et/4 + 3et/2 6et/4 − 6et/2
−et/4 + et/2 3et/4 − 2et/2
) (
c1
c2
)
= c1
(−2
−1
)
et/4 + c1
(
3
1
)
et/2 + c2
(
6
3
)
et/4 + c2
(−6
−2
)
et/2
= (−c1 + 3c2)
(
2
1
)
et/4 + (c1 − 2c2)
(
3
1
)
et/2
= c3
(
2
1
)
et/4 + c4
(
3
1
)
et/2.
23. From equation (3) in the text we have eDt = I + tD +
t2
2!
D2 +
t3
3!
D3 + · · · so that
PeDtP−1 = PP−1 + t(PDP−1) +
t2
2!
(PD2P−1) +
t3
3!
(PD3P−1) + · · · .
597
10.5 Matrix Exponential
But PP−1 + I, PDP−1 = A and PDnP−1 = An (see Problem 37, Exercises 8.12). Thus,
PeDtP−1 = I + tA +
t2
2!
A2 +
t3
3!
A3 + · · · = eAt.
24. From equation (3) in the text
eDt =

1 0 · · · 0
0 1 · · · 0
...
...
. . .
...
0 0 · · · 1
 +

λ1 0 · · · 0
0 λ2 · · · 0
...
...
. . .
...
0 0 · · · λn
 + 12! t2

λ21 0 · · · 0
0 λ22 · · · 0
...
...
. . .
...
0 0 · · · λ2n

+
1
3!
t3

λ31 0 · · · 0
0 λ32 · · · 0
...
...
. . .
...
0 0 · · · λ3n
 + · · ·
=

1 + λ1t + 12! (λ1t)
2 + · · · 0 · · · 0
0 1 + λ2t + 12! (λ2t)
2 + · · · · · · 0
...
...
. . .
...
0 0 · · · 1 + λnt + 12! (λnt)2 + · · ·

=

eλ1t 0 · · · 0
0 eλ2t · · · 0
...
...
. . .
...
0 0 · · · eλnt

25. From Problems 23 and 24 and equation (1) in the text
X = eAtC = PeDtP−1C =
(
e3t e5t
e3t 3e5t
) (
e3t 0
0 e5t
) ( 3
2e
−3t − 12e−3t
− 12e−5t 12e−5t
)(
c1
c2
)
=
(
3
2e
3t − 12e5t − 12e3t + 12e5t
3
2e
3t − 32e5t − 12e3t + 32e5t
)(
c1
c2
)
.
26. From Problems 23 and 24 and equation (1) in the text
X = eAtC = PeDtP−1C =
(−et e3t
et e3t
) (
et 0
0 e3t
) (− 12e−t 12e−t
1
2e
3t 1
2e
−3t
)(
c1
c2
)
=
(
1
2e
t + 12e
9t − 12et + 12e3t
− 12et + 12e9t 12et + 12e3t
)(
c1
c2
)
.
27. (a) The following commands can be used in Mathematica:
A={{4, 2},{3, 3}};
c={c1, c2};
m=MatrixExp[A t];
sol=Expand[m.c]
Collect[sol, {c1, c2}]//MatrixForm
598
10.5 Matrix Exponential
The output gives
x(t) = c1
(
2
5
et +
3
5
e6t
)
+ c2
(
−2
5
et +
2
5
e6t
)
y(t) = c1
(
−3
5
et +
3
5
e6t
)
+ c2
(
3
5
et +
2
5
e6t
)
.
The eigenvalues are 1 and 6 with corresponding eigenvectors(−2
3
)
and
(
1
1
)
,
so the solution of the system is
X(t) = b1
(−2
3
)
et + b2
(
1
1
)
e6t
or
x(t) = −2b1et + b2e6t
y(t) = 3b1et + b2e6t.
If we replace b1 with − 15c1 + 15c2 and b2 with 35c1 + 25c2, we obtain the solution found using the matrix
exponential.
(b) x(t) = c1e−2t cos t− (c1 + c2)e−2t sin t
y(t) = c2e−2t cos t + (2c1 + c2)e−2t sin t
28. x(t) = c1(3e−2t − 2e−t) + c3(−6e−2t + 6e−t)
y(t) = c2(4e−2t − 3e−t) + c4(4e−2t − 4e−t)
z(t) = c1(e−2t − e−t) + c3(−2e−2t + 3e−t)
w(t) = c2(−3e−2t + 3e−t) + c4(−3e−2t + 4e−t)
29. If det(sI−A) = 0, then s is an eigenvalue of A. Thus sI−A has an inverse if s is not an eigenvalue of A. For
the purposes of the discussion in this section, we take s to be larger than the largest eigenvalue of A. Under
this condition sI−A has an inverse.
30. Since A3 = 0, A is nilpotent. Since
eAt = I + At + A2
t2
2!
+ · · ·+ Ak t
k
k!
+ · · · ,
if A is nilpotent and Am = 0, then Ak = 0 for k ≥ m and
eAt = I + At + A2
t2
2!
+ · · ·+ Am−1 t
m−1
(m− 1)! .
In this problem A3 = 0, so
eAt = I + At + A2
t2
2
=
 1 0 00 1 0
0 0 1
 +
−1 1 1−1 0 1
−1 1 1
 t +
−1 0 10 0 0
−1 0 1
 t22
=
 1− t− t
2/2 t t + t2/2
−t 1 t
−t− t2/2 t 1 + t + t2/2

and the solution of X′ = AX is
X(t) = eAtC = eAt
 c1c2
c3
 =
 c1(1− t− t
2/2) + c2t + c3(t + t2/2)
−c1t + c2 + c3t
c1(−t− t2/2) + c2t + c3(1 + t + t2/2)
 .
599
10.5 Matrix Exponential
CHAPTER 10 REVIEW EXERCISES
CHAPTER 10 REVIEW EXERCISES
1. If X = k
(
4
5
)
, then X′ = 0 and
k
(
1 4
2 −1
) (
4
5
)
−
(
8
1
)
= k
(
24
3
)
−
(
8
1
)
=
(
0
0
)
.
We see that k = 13 .
2. Solving for c1 and c2 we find c1 = − 34 and c2 = 14 .
3. Since  4 6 61 3 2
−1 −4 −3

 31
−1
 =
 124
−4
 = 4
 31
−1
 ,
we see that λ = 4 is an eigenvalue with eigenvector K3. The corresponding solution is X3 = K3e4t.
4. The other eigenvalue is λ2 = 1− 2i with corresponding eigenvector K2 =
(
1
−i
)
. The general solution is
X(t) = c1
(
cos 2t
− sin 2t
)
et + c2
(
sin 2t
cos 2t
)
et.
5. We have det(A− λI) = (λ− 1)2 = 0 and K =
(
1
−1
)
. A solution to (A− λI)P = K is P =
(
0
1
)
so that
X = c1
(
1
−1
)
et + c2
[(
1
−1
)
tet +
(
0
1
)
et
]
.
6. We have det(A− λI) = (λ + 6)(λ + 2) = 0 so that
X = c1
(
1
−1
)
e−6t + c2
(
1
1
)
e−2t.
7. We have det(A− λI) = λ2 − 2λ + 5 = 0. For λ = 1 + 2i we obtain K1 =
(
1
i
)
and
X1 =
(
1
i
)
e(1+2i)t =
(
cos 2t
− sin 2t
)
et + i
(
sin 2t
cos 2t
)
et.
Then
X = c1
(
cos 2t
− sin 2t
)
et + c2
(
sin 2t
cos 2t
)
et.
8. We have det(A− λI) = λ2 − 2λ + 2 = 0. For λ = 1 + i we obtain K1 =
(
3− i
2
)
and
X1 =
(
3− i
2
)
e(1+i)t =
(
3 cos t + sin t
2 cos t
)
et + i
(− cos t + 3 sin t
2 sin t
)
et.
600
CHAPTER 10 REVIEW EXERCISES
Then
X = c1
(
3 cos t + sint
2 cos t
)
et + c2
(− cos t + 3 sin t
2 sin t
)
et.
9. We have det(A− λI) = −(λ− 2)(λ− 4)(λ + 3) = 0 so that
X = c1
−23
1
 e2t + c2
 01
1
 e4t + c3
 712
−16
 e−3t.
10. We have det(A−λI) = −(λ+2)(λ2−2λ+3) = 0. The eigenvalues are λ1 = −2, λ2 = 1+
√
2i, and λ2 = 1−
√
2i,
with eigenvectors
K1 =
−75
4
 , K2 =
 11
2
√
2 i
1
 , and K3 =
 1− 12√2 i
1
 .
Thus
X = c1
−75
4
 e−2t + c2

 10
1
 cos√2t−
 012√2
0
 sin√2t
 et
+ c3

 012√2
0
 cos√2t +
 10
1
 sin√2t
 et
= c1
−75
4
 e−2t + c2
 cos
√
2t
− 12
√
2 sin
√
2t
cos
√
2t
 et + c3
 sin
√
2t
1
2
√
2 cos
√
2t
sin
√
2t
 et.
11. We have
Xc = c1
(
1
0
)
e2t + c2
(
4
1
)
e4t.
Then
Φ =
(
e2t 4e4t
0 e4t
)
, Φ−1 =
(
e−2t −4e−2t
0 e−4t
)
,
and
U =
∫
Φ−1F dt =
∫ (
2e−2t − 64te−2t
16te−4t
)
dt =
(
15e−2t + 32te−2t
−e−4t − 4te−4t
)
,
so that
Xp = ΦU =
(
11 + 16t
−1− 4t
)
.
12. We have
Xc = c1
(
2 cos t
− sin t
)
et + c2
(
2 sin t
cos t
)
et.
Then
Φ =
(
2 cos t 2 sin t
− sin t cos t
)
et, Φ−1 =
(
1
2 cos t − sin t
1
2 sin t cos t
)
e−t,
and
U =
∫
Φ−1F dt =
∫ (
cos t− sec t
sin t
)
dt =
(
sin t− ln | sec t + tan t|
− cos t
)
,
601
CHAPTER 10 REVIEW EXERCISES
so that
Xp = ΦU =
( −2 cos t ln | sec t + tan t|
−1 + sin t ln | sec t + tan t|
)
et.
13. We have
Xc = c1
(
cos t + sin t
2 cos t
)
+ c2
(
sin t− cos t
2 sin t
)
.
Then
Φ =
(
cos t + sin t sin t− cos t
2 cos t 2 sin t
)
, Φ−1 =
(
sin t 12 cos t− 12 sin t
− cos t 12 cos t + 12 sin t
)
,
and
U =
∫
Φ−1F dt =
∫ ( 1
2 sin t− 12 cos t + 12 csc t
− 12 sin t− 12 cos t + 12 csc t
)
dt
=
(− 12 cos t− 12 sin t + 12 ln | csc t− cot t|
1
2 cos t− 12 sin t + 12 ln | csc t− cot t|
)
,
so that
Xp = ΦU =
(−1
−1
)
+
(
sin t
sin t + cos t
)
ln | csc t− cot t|.
14. We have
Xc = c1
(
1
−1
)
e2t + c2
[(
1
−1
)
te2t +
(
1
0
)
e2t
]
.
Then
Φ =
(
e2t te2t + e2t
−e2t −te2t
)
, Φ−1 =
(−te−2t −te−2t − e−2t
e−2t e−2t
)
,
and
U =
∫
Φ−1F dt =
∫ (
t− 1
−1
)
dt =
( 1
2 t
2 − t
−t
)
,
so that
Xp = ΦU =
(− 12
1
2
)
t2e2t +
(−2
1
)
te2t.
15. (a) Letting
K =
 k1k2
k3

we note that (A− 2I)K = 0 implies that 3k1 + 3k2 + 3k3 = 0, so k1 = −(k2 + k3). Choosing k2 = 0, k3 = 1
and then k2 = 1, k3 = 0 we get
K1 =
−10
1
 and K2 =
−11
0
 ,
respectively. Thus,
X1 =
−10
1
 e2t and X2 =
−11
0
 e2t
602
CHAPTER 10 REVIEW EXERCISES
are two solutions.
(b) From det(A− λI) = λ2(3− λ) = 0 we see that λ1 = 3, and 0 is an eigenvalue of multiplicity two. Letting
K =
 k1k2
k3
 ,
as in part (a), we note that (A − 0I)K = AK = 0 implies that k1 + k2 + k3 = 0, so k1 = −(k2 + k3).
Choosing k2 = 0, k3 = 1, and then k2 = 1, k3 = 0 we get
K2 =
−10
1
 and K3 =
−11
0
 ,
respectively. Since the eigenvector corresponding to λ1 = 3 is
K1 =
 11
1
 ,
the general solution of the system is
X = c1
 11
1
 e3t + c2
−10
1
 + c3
−11
0
 .
16. For X =
(
c1
c2
)
et we have X′ = X = IX.
603

Outros materiais