Download as pdf or txt
Download as pdf or txt
You are on page 1of 210

Differential Equations and Linear Algebra

2nd Edition Farlow


Full download at link: https://testbankpack.com/p/solution-manual-
for-differential-equations-and-linear-algebra-2nd-edition-by-farlow-
isbn-0134689542-9780134689548/

CHAPTER Higher-Order
Second-Order Linear
Linear
4 Differential Equations
Differential Equations

4.1 The Harmonic Oscillator

 The Undamped Oscillator


1. x + x = 0, x (0) = 1 , x (0) = 0

The general solution of the harmonic oscillator equation x + x = 0 is given by


x ( t ) = c1 cost + c2 sin t
x ( t ) = −c1 sin t + c2 cost .

Substituting the initial conditions x ( 0 ) = 1 , x ( 0 ) = 0 , gives

x ( 0 ) = c1 = 1
x ( 0 ) = c2 = 0

so c1 =1, c2 = 0 . Hence, the IVP has the solution x ( t ) = cost .

2. x + x = 0, x (0) = 1 , x (0) = 1

The general solution of the harmonic oscillator equation x + x = 0 is given by


x ( t ) = c1 cost + c2 sin t
x ( t ) = −c1 sin t + c2 cost .

Substituting the initial conditions x ( 0 ) = 1 , x ( 0 ) = 1 , gives

x ( 0 ) = c1 = 1
x ( 0 ) = c2 = 1

or c1 = c2 = 1. Hence, the IVP has the solution


x ( t ) = cos t + sin t .

294
295 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 295

In polar form, this would be


⎛ π⎞
x ( t ) = 2 cos t − .
⎜ ⎟
⎝ 4⎠
3. x + 9x = 0 , x (0) = 1 , x (0) = 1

The general solution of the harmonic oscillator equation x + 9x = 0 is given by


x ( t ) = c1 cos3t + c2 sin 3t
x ( t ) = −3c1 sin 3t + 3c2 cos 3t .

Substituting the initial conditions x ( 0 ) = 1 , x ( 0 ) = 1 , gives

x ( 0 ) = c1 = 1
x ( 0 ) = 3c2 = 1

1
so c1 = 1, c2 = . Hence, the IVP has the solution
3
1
x ( t ) = cos 3t + sin 3t .
3
In polar form, this would be
10
x (t ) = cos ( 3t − δ )
3
1
where δ = tan −1 . This would be in the first quadrant.
3
4. x + 4x = 0 , x ( 0 ) = 1 , x ( 0 ) = −2

The general solution of the harmonic oscillator equation x + 4x = 0 is given by


x ( t ) = c1 cos 2t + c2 sin 2t
x ( t ) = −2c1 sin 2t + 2c2 cos 2t .

Substituting the initial conditions x ( 0 ) = 1 , x ( 0 ) = −2 , gives

x ( 0 ) = c1 = 1
x ( 0 ) = 2c2 = −2

so c1 = 1, c2 = −1 . Hence, the IVP has the solution


x ( t ) = cos 2t − sin 2t .
In polar form, this would be
⎛ π⎞
x ( t ) = 2 cos 2t + .
⎜ ⎟
⎝ 4⎠
296 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 296

5. x + 16x = 0 , x ( 0 ) = −1 , x ( 0 ) = 0

The general solution of the harmonic oscillator equation x + 16x = 0 is given by


x ( t ) = c1 cos 4t + c2 sin 4t
x ( t ) = −4c1 sin 4t + 4c2 cos 4t .

Substituting the initial conditions x ( 0 ) = −1 , x ( 0 ) = 0 , gives

x ( 0 ) = c1 = −1
x ( 0 ) = 4c2 = 0

so c1 = −1, c2 = 0 . Hence, the IVP has the solution


x ( t ) = − cos 4t .

6. x + 16x = 0 , x (0) = 0 , x (0) = 4

The general solution of the harmonic oscillator equation x + 16x = 0 is given by


x ( t ) = c1 cos 4t + c2 sin 4t
x ( t ) = −4c1 sin 4t + 4c2 cos 4t .

Substituting the initial conditions x ( 0 ) = 0 , x ( 0 ) = 4 , we get

x ( 0 ) = c1 = 0
x ( 0 ) = 4c2 = 4

so c1 = 0 , c2 = 1. The IVP has the solution

x ( t ) = sin 4t .

7. x + 16π 2 x = 0 , x (0) = 0, x (0) = π


16π 2
ω0 = = 4π
1

x = c1 cos 4π t + c2 sin 4π t

x = −4π c1 sin 4π t + 4π c2 cos 4π t

x(0) = 0 = c1
1
x(0) = π = 4π c2 c2 =
4
1
x= sin 4π t
4
297 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 297

8. 4 x + π 2 x = 0 , x (0) = 1, x ( 0 ) = π

π2 π
ω0 = =
4 2

x = c1 cos 4π t + c2 sin 4π t
π π π π
x = c1 sin t + c2 cos t
2 2 2 2

x(0) = 1 = c1
π
x(0) = π = c2 , c2 = 2
2

π π
x = cos t + 2sin t
2 2

 Graphing by Calculator

9. y = cos t + sin t y
1.5 T = 2π
The equation tells us T = 2π and because


T= , ω0 = 1 . We then measure the delay
ω0
δ
≈ 0.8 which we can compute as the phase π t
ω0 δ A ≈ 1.4
≈ 0.8
angle δ ≈ 0.8 (1) = 0.8 . The amplitude A can be ω0

measured directly giving A ≈ 1.4 . Hence,


Š1.5
cost + sin t ≈ 1.4cos ( t − 0.8 ) .
Compare with the algebraic form in Problem 15.

10. y = 2cost + sin t y

2.5
The equation tells us T = 2π and because T = 2π
δ A ≈ 2.2
2π ≈ 0.5
T= , ω0 = 1 . We then measure the delay ω0
ω0
δ t
≈ 0.5 , which we can compute as the phase Š4 4 8
ω0
angle δ ≈ 0.5 (1) = 0.5 . The amplitude A can be

measured directly giving A ≈ 2.2 . Hence, Š2.5


298 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 298

2cost + sin t ≈ 2.2cos ( t − 0.5 ) .


299 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 299

11. y = 5cos3t + sin 3t yδ


≈ 0.05
ω0
2π 5
T = 2π /3 A ≈ 5.1
The equation tells us that period is T = and
3


because T = , ω0 = 3 . We then measure the
ω0
t
δ π
delay ≈ 0.05 , which we can compute as the
ω0

phase angle δ ≈ 3 ( 0.05 ) = 0.15 . The amplitude A


–5
can be measured directly giving A ≈ 5.1 . Hence,

5cos3t + sin 3t ≈ 5.1cos ( 3t − 0.15 ) .

12. y = cos 3t + 5sin 3t y δ


≈ 0.5
ω0
2π 5
T = 2π /3 A ≈ 5.1
The equation tells us the period is T = and
3


because T = , ω0 = 3 . We then measure the
ω0
t
δ 3
delay ≈ 0.5 , which we can compute as the
ω0

phase angle δ ≈ 0.5 ( 3) = 1.5 . The amplitude A


Š5
can be measured directly giving A ≈ 5.1 . Hence,

cos3t + 5sin 3t ≈ 5.1cos ( 3t −1.5 ) .

13. y = − cos 5t + 2sin 5t yδ


≈ 0.4
ω0
2π 2 T = 2π /5 A ≈ 2.2
equation tells us that period is T = and
5


because T = , ω0 = 5 . We then measure the
ω0 1 2
t

δ π
delay ≈ or 0.4 , which we can compute as
ω0 8
–2
the phase angle δ ≈ 5 ( 0.4 ) = 2 . The amplitude A
can be measured directly giving A ≈ 2.2 . Hence,

− cos 5t + 2sin t ≈ 2.2cos ( 5t − 2 ) .


300 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 300

 Alternate Forms for Sinusoidal Oscillations

14. We have
Acos (ω0t − δ ) = A ( cos ω0t cos δ + cos ω0 t cos δ ) = ( Acos δ ) cos ω0 t + ( Asin δ ) sin ω0 t
= c1 cos ω0 t + c2 sin ω0t

where c1 = Acos δ , c2 = Asin δ .

 Single-Wave Forms of Simple Harmonic Motion

15. cost + sin t


By Equation (4) c1 = 1, c2 = 1, and ω0 = 1 . By Equation (5)
π
A= 2 , δ =
4
yielding
⎛ π⎞
cost + sin t = 2 cos t − .
⎜ ⎟
4⎠

(Compare with solution to Problem 9.)

16. cost − sin t


By Equation (4) c1 = 1, c2 = −1 , and ω0 = 1 . By Equation (5)
π
A= 2 , δ =−
4
yielding
π
cost − sin t = 2 cos ⎛ t + ⎞ .
⎜ 4 ⎟⎠

Because c1 is positive and c2 is negative the phase angle is in the 4th quadrant.
17. − cost + sin t
By Equation (4) c1 = −1, c2 = 1, and ω0 = 1 . By Equation (5)

A= 2 , δ =
4
yielding
⎛ 3π ⎞
− cos t + sin t = 2 cos t − .
⎜ ⎟
4 ⎠

Because c1 is negative and c2 is positive the phase angle is in the 2nd quadrant.
301 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 301

18. − cost − sin t


By Equation (5) c1 = −1, c2 = −1 , and ω0 = 1 . By Equation (6)

A= 2 , δ =
4

yielding

⎛ 5π ⎞
− cos t − sin t = 2 cos t − .
⎜ ⎟
⎝ 4 ⎠
Because c1 and c2 are negative, the phase angle is in the 3rd quadrant.

 Component Form of Harmonic Motion


Using cos ( A + B ) = cos Acos B − sin Asin B , we write:

19. 2cos ( 2t − π ) = 2 {cos 2t cos ( −π ) − sin 2t sin ( −π )} = −2 cos 2t

⎛ π⎞ ⎛π ⎞ ⎛π⎞ 1 3
20. cos t + = cost cos − sin t sin = cost − sin t
⎜ 3 ⎟⎠ ⎜3⎟ ⎜3⎟ 2 2
⎝ ⎝ ⎠ ⎝ ⎠
⎛ π⎞ ⎧ ⎛ π⎞ ⎛ π ⎞⎫ 2 2 3 2
21. 3cos t − = 3 cos t cos − − sin t sin − =3 cos t + sin t = {cos t + sin t}
⎜ ⎟ ⎨ ⎜ ⎟ ⎜ ⎟⎬ ⎨ ⎬
4 4 4 2 2 2
⎝ ⎠ ⎩ ⎝ ⎠ ⎝ ⎠⎭

⎛ π⎞ ⎛ π⎞ ⎛ π⎞ 3 1
22. cos 3t − = cos3t cos − − sin 3t sin − = cos 3t + sin 3t
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
6 6 6 2 2
⎝ ⎠ ⎝ ⎠ ⎝ ⎠
 Interpreting Oscillator Solutions
23. x + x = 0 , x (0) = 1 , x (0) = 0
1
Because ω0 = 1 , we know the natural frequency is Hz and the period is 2π seconds. Using

the initial conditions, we find the solution (see Problem 1)


x ( t ) = cost ,

which tells us the amplitude is 1 and the phase angle δ = 0 radians.


24. x + x = 0 , x (0) = 1 , x (0) = 1
1
Because ω0 = 1 radians per second, we know the natural frequency is Hz (cycles per

second), and the period is 2π . Using the initial conditions, we find the solution (see Problem 2)
⎛ π⎞
x ( t ) = 2 cos t − ,
⎜ 4 ⎟
⎝ ⎠
302 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 302

π
which tells us the amplitude is 2 and the phase angle is δ =
4
radians.
301 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 301

25. x + 9x = 0 , x ( 0 ) = 1 , x ( 0 ) = 1
3
Because ω0 = 3 radians per second, we know the natural frequency is Hz (cycles per


second), and the period is . Using the initial conditions, we find the solution (see Problem 3)
3
10
x (t ) = cos ( 3t − δ )
3
1
where δ = tan −1 , which tells us the amplitude is
3
10
3
and the phase angle is
1
δ = tan −1 ≈ 0.3218 radians.
3
26. x + 4x = 0 , x ( 0 ) = 1 , x ( 0 ) = −2

1
Because ω0 = 2 radians per second, we know the natural frequency is Hz (cycles per second),
π

and the period is π. Using the initial conditions, we find the solution (see Problem 4)
⎛ π⎞
x ( t ) = 2 cos 2t + ,
⎜ ⎟
4⎠

π
which tells us the amplitude is 2 and the phase angle is δ = − radians.
4

27. x + 16x = 0 , x ( 0 ) = −1 , x ( 0 ) = 0

2
Because ω0 = 4 radians per second, we know the natural frequency is Hz (cycles per second),
π
π
and the period is . Using the initial conditions, we find the solution (see Problem 5)
2
x ( t ) = cos ( 4t − π ) ,
which tells us the amplitude is 1 and the phase angle is δ = π radians.
302 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 302

28. x + 16x = 0 , x ( 0 ) = 0 , x ( 0 ) = 4
2
Because ω0 = 4 radians per second, we know the natural frequency is Hz (cycles per second),
π
π
and the period is . Using the initial conditions, we find the solution (see Problem 6)
2
⎛ π⎞
x ( t ) = cos ⎜ 4t − ⎟ ,
⎝ 2⎠
π
which tells us the amplitude is 1 and the phase angle is δ = radians.
2

29. x + 16π 2 x = 0 , x (0) = 0, x ( 0 ) = π

1
From Problem 7, x = sin 4π t
4
1 1 ⎛ π⎞ π 2π 2
Amplitude = ,x= cos 4π t − , phase angle δ = , and period T = = = 8.
4 4 ⎜ ⎟
2⎠ 2 ω π
⎝ 0
4
30. 4 x + π 2 x = 0 , x (0) = 1, x ( 0 ) = π
4r2 + π2 = 0

π π π
r= ± i x = c cos t + c sin t 1=c
1 2 1
2 2 2
π π π π π
x=− c1 sin t + c2 cos t π= c
2 2 2 2 2 2

π π
x = cos t + 2sin t c2 = 2
2 2

Amplitude: A = 1+ 22 = 5

⎛π ⎞
x= 5 cos ⎜ t −1.11 ⎟
⎝2 ⎠

 Relating Graphs

31. (a) See graph, next page.


k
(b) x + 0.25x = 0 ω0 = = 0.5
m

From (4) x = c1 cos 0.5t + c2 sin 0.5t


⎛t⎞ c ⎛t⎞
303 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 303

x(0) = 0 ⇒ c1 = 0 so x(t) = c2 sin ⎜ ⎟ and x(t) = 2 cos ⎜ ⎟ .


⎝2⎠ 2 ⎝2⎠
304 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 304

Alternatively, you could use (5)

x = A cos (0.5t − δ)
π ⎛t π⎞ ⎛t ⎞
x(0) = 0 ⇒ δ = so x(t) = A cos ⎜ − ⎟ = A sin ⎜ ⎟ ,
2 ⎝2 2⎠ ⎝2⎠
A ⎛t⎞
and x(t) = cos .
2 ⎜ 2⎟
⎝ ⎠

(c) See graph

Graph for b) and d) Graph for a)


A A 2A 5A
(d) Amplitudes are approximately , , , , and A
3 2 3 6

 Phase Portraits

For comparison of phase portraits, the main observation is that the elliptical shape depends on ω0,
which is k in all of these problems because x + kx = 0 .

If ω0 = 1, trajectories are circular. As ω0 increases above 1, ellipses become taller and thinner.
As ω0 decreases from 1 to 0, ellipses become shorter and wider.
x 1
The aspect ratio of max = .
xmax ω

Other observations include:


• All these phase portraits show closed elliptical trajectories that circulate clockwise.
• The trajectory of Problem 33 has a greater radius than that of Problem 32 because the initial
condition is further from the origin.
• The trajectories in Problems 36 and 37 are on the same ellipse with different starting points
that give different solution equations.
305 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 305

32. G ⎡1 ⎤
x+ x=0 x0 = ⎢ ⎥
⎣ 0⎦

From Problem 1 x(t) = cos(t),


so x(t) = −sin(t).

33. G ⎡1⎤
x+ x=0 x0 = ⎢ ⎥
⎣1⎦

From Problem 2 x(t) = cos(t) + sin(t),


so x(t) = −sin(t) + cos(t).

34. G ⎡1⎤
x + 9x = 0 x0 = ⎢ ⎥
⎣1⎦
1
From Problem 3, x(t) = cos 3t + sin 3t,
3
so x(t) = −3 sin 3t + cos 3t.
306 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 306

35. G ⎡1⎤
x + 4x = 0 x0 = ⎢ ⎥
⎣ −2⎦

From Problem 4, x(t) = cos 2t − sin 2t,


so x(t) = −2 sin 2t − 2 cos 2t.

36. G ⎡ −1⎤
x + 16x = 0 x0 = ⎢ ⎥
⎣0⎦

From Problem 5, x(t) = −cos 4t,


so x′(t) = 4 sin 4t.

37. G ⎡ 0⎤
x + 16x = 0 x0 = ⎢ ⎥
⎣ 4⎦

From Problem 6, x(t) = sin 4t,


so x(t) = 4 cos 4t.
306 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 306

38. G ⎡0 ⎤
x + 16π 2 x = 0 x0 = ⎢ ⎥
⎣π ⎦
1
From Problem 7, x(t) = sin 4π t,
4
so x(t) = cos 4π t.

39. G ⎡1 ⎤
4x + π 2 x = 0 x0 = ⎢ ⎥
⎣π ⎦
π π
From Problem 8, x(t) = cos t + 2sin t,
2 2
π π π
so x(t) = − sin t + π cos t.
2 2 2

 Matching Problems

40. B

41. A

42. D

43. C
307 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 307

 Changing Frequencies
(a) ω0 = 0.5 gives tx curve with lowest fre- x ω 0 = 0.5 ω = 2
44. 4 0

quency (fewest humps); ω0 = 2 gives the

highest frequency (most humps). 2


(b) ω0 = 0.5 gives the innermost phase-

plane trajectory; as ω0 increases, the t


π 2π 3

amplitude of x increases. In Figure 4.1.8


–2
the trajectory that is not totally visible is
the one for ω0 = 2 .
–4 ω0 = 1

 Detective Work
⎛ 8π ⎞
45. (a) The curve y = 1.4cos t − is a sinusoidal curve with period 2π , amplitude A ≈ 1.4 ,
⎜ ⎟
5 ⎠


and phase angle δ ≈ .
5
π
(b) From this graph we estimate ω0 = 1 , A ≈ 2.3 , and δ ≈ . Thus, we have
4
⎛ π⎞ ⎡ ⎛ π⎞ ⎛ π ⎞⎤
x ( t ) = Acos (ω0t − δ 0 ) = 2.3cos ⎜ t − ⎟ = 2.3 ⎢ cost ⎜ cos− ⎟ − sin t ⎜ −sin ⎟⎥
⎝ 4⎠ ⎣ ⎝ 4⎠ ⎝ 4 ⎠⎦

⎧⎪ 2 2 ⎪⎫
= 2.3 ⎨ cos t + sin t ⎬ ≈ 1.6 ( cos t + sin t ) .
⎪⎩ 2 2 ⎪⎭

 Pulling a Weight
46. (a) The mass is m = 2 kg . Because a force of 8 nt stretches the spring 0.5 meters, we find

8
that k = = 16 nt m . If we then release the weight, the IVP describing the motion of
0.5

the weight is 2 x +16x = 0 or


x + 8x = 0 , x ( 0 ) = 0.5 , x ( 0 ) = 0 .

The solution of the differential equation is

x ( t ) = Acos ( 8t − δ . )
308 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 308

Using the initial conditions, we get the simple oscillation

x ( t ) = 0.5cos ( 8t ) .
309 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 309

Amplitude =
1 2π 2π 8
(b) m; T = = sec , f = cycles per second
2 ω0 8 2π

)
(c) Setting cos ( 8t = 0 , we find that the weight will pass through equilibrium at
1
4
of the

period or after
π
t= ≈ 0.56 seconds.
2 8

At that time velocity is


⎛π⎞
x ( 0.56 ) = − 2 sin ⎜ ⎟ ≈ −1.414m sec
⎝2⎠

moving away from original displacement.

 Finding the Differential Equation


47. (a) The mass is m = 500 gm, which means the force acting on the spring is 500 × 980 dynes.

This stretches the spring 50 cm, so the spring constant is


500 ×980
k= = 9800 dynes cm .
50
The mass is then pulled down 10 cm from its initial displacement, giving x ( 0 ) = 10 (as

long as we measure downward to be the positive direction, which is typical in these


problems). The initial velocity of the mass is assumed to be zero, so x ( 0 ) = 0 . Thus, the
IVP for the mass is

500 x + 9800x = 0
or
5 x + 98x = 0 , x ( 0 ) = 10 , x ( 0 ) = 0 .

(b) The solution of the differential equation found in part (a) is


⎛ 98 ⎞ 98
x ( t ) = Acos t − δ = 10cos t.
⎜ 5 ⎟ 5
⎝ ⎠

(c) In part (b) the amplitude is 10 cm, phase angle is 0, the period is
m 5
T = 2π = 2π ≈ 1.4 sec,
k 98
310 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 310

and the natural frequency is given by the reciprocal


1
f = = 0.71 oscillations per sec-
ond. 1.4
309 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 309

 Initial-Value Problems
16 1
48. (a) The weight is 16 lbs, so the mass is roughly = slugs. (See Table 4.1.1 in text.) This
32 2
1 16
mass stretches the spring foot, hence k = 1 = 32 lb ft . This yields the equation
2 2
1
2
( x ) + 32x = 0 , or

x + 64x = 0 .
1
The initial conditions are that the mass is pulled down 4 inches ( foot) from equilib-
3

rium and then given an upward velocity of 4 ft sec . This gives the initial conditions of
1
x (0) = ft, x ( 0 ) = −4 ft/sec, using the engineering convention that for x, down is
3
positive.

1
(b) We have the same equation x + 64x = 0 , but the initial conditions are x ( 0 ) = − ft,
6
x ( 0 ) = 1 ft/sec.

 One More Weight

12 3 1
49. The mass is m = = slugs. The spring is stretched foot, so the spring constant is
32 8 2

12 1 1
k= = 24 lb ft . The initial position of the mass is 4 inches ( ft) upward so x ( 0 ) = − 3 . The
3
1
2

initial motion is 2 ft sec upward, and thus x ( 0 ) = −2 . Hence, the equation for the motion of the
mass is

1
x + 64x = 0 , x ( 0 ) = − , x ( 0 ) = −2 ,
3

which has the solution


1 1
x ( t ) = − cos8t − sin 8t .
3 4
310 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 310

Writing this in polar form, we have x

2 2 0.4
⎛ 1⎞ ⎛ 1⎞ 5
A = c2 + c2 = − + − =
1 2 ⎜ 3⎟ ⎜ 4⎟ 12
⎝ ⎠ ⎝ ⎠
⎛c ⎞ ⎛3⎞
δ = tan −1 ⎜ 2 ⎟ = tan −1 ⎜ ⎟ t
c
⎝ 1⎠ ⎝4⎠ 0.2 0.4 0.6 0.8 1 1.2 1.4

≈ 3.78 radians ( angle in 3rd quadrant ) .


Hence, we have the solution in polar form Š0.4 5
x(t) = cos(8t − 3.78)
5 12
x (t ) = cos ( 8t − 3.78 ) .
12
Spring oscillation
See figure.

 Comparing Harmonic Motions


2π k
50. The period of simple harmonic motion is given by T = , where ω0 = . Notice that this
ω0 m

does not depend at all on our initial conditions. Period is the same so is the frequency, but the
amplitude will be twice that in the first case.

 Testing Your Intuition


51. x + x + x3 = 0

Here we have a vibrating spring with no friction, but a nonlinear restoring force F = −x − x3 that

is stronger than a purely linear force –x. For small displacement x the nonlinear F will not be
much different (for small x, x 3 is very small), but for larger x, the force F will be much stronger

than in a linear spring; as F increases, the frequency of the vibration increases. This equation is
called Duffing’s (strong) equation, and the associated springs are called strong springs.

52. x + x − x3 = 0

Here we have a vibrating spring with no friction, and a nonlinear restoring force F = −x + x3 . For

small displacement x the nonlinear term x 3 has little effect, but as x increases toward 1, the

restoring force F diminishes (i.e., the spring weakens when it is stretched a lot, and the restoring
force becomes zero when x = 1 ). The decreasing F causes decreasing frequency (and increasing

period). This equation is called Duffing’s (weak) equation, and the associated springs are called
weak springs.
311 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 311

53. x−x=0
This equation describes a spring with no friction and a negative restoring force. You may wonder
if there are such physical systems. In the next two sections we will see that this equation describes
the motion of an inverted pendulum (4.3 Problems 58, 59), and it has solutions sinh t and cosh t
(4.2 Example 2), in contrast to x + x = 0 , which has solutions sin t and cos t. The restoring force

for the equation under discussion is always directed away from the equilibrium position; hence
the solution always moves away from the equilibrium, which us unstable.
1
54. x+ x+x=0
t
This equation can be interpreted as describing the motion of a vibrating mass that has infinite
1
friction x at t = 0 , but friction immediately begins to diminish and approaches zero as t
t

becomes very large. You may simulate in your mind the motion of such a system. Do you think for
large t that the oscillation might behave much like simple harmonic motion? (See 4.3 Problem 68.)
55. ( )
x + x 2 −1 x + x = 0

This is called van der Pol’s equation and describes oscillations (mostly electrical) where internal
friction depends on the value of the dependent variable x. Note that when x < 1, we actually have

negative friction, so for a small displacement x we would expect the system to move away from
the zero solution (an unstable equilibrium) in the direction of x = 1. But when x > 1, we will

have positive friction causing damping. We will see in 4.3 Problem 70 and in Chapter 7 that there
is a periodic solution between small x and large x that attracts all these other solutions.

56. x + tx = 0
Here we have a vibrating spring with no friction, but the restoring force –tx gets stronger as time
passes. Hence we expect to see no damping, but faster vibrations as t increases.

 LR-Circuit

57. (a) Without having a capacitor to store energy, we do not expect the current in the circuit to
oscillate. If there had been a constant voltage V0 on in the past, we would expect the

V0
current to be (by Ohm’s law) I = . If we then shut off the voltage, we would expect
R

the current to die off in the presence of a resistance.

(b) If a current I passes through a resistor with resistance R, then the voltage drop is RI; the
voltage drop across an inductor of inductance L is LI . We obtain the IVP:
312 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 312

V0
LI + RI = 0 , I ( 0 ) =
R
313 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 313

(c) The solution of the IVP is


V0 −( R L )t
I (t ) = e .
R
1
(d) If R = 40 ohms, L = 5 henries, V0 = 10 volts, then I ( t ) = e −8t ohms.
4

 LC-Circuit

58. (a) With a nonzero initial current and no resistance, we do not expect the current to damp to
zero. We would expect an oscillatory current due to the capacitor. Thus the charge on the
capacitor would oscillate indefinitely. The exact behavior depends on the initial
conditions and the values of the inductance and capacitance.

(b) Kirchoff’s voltage law states that the sum of the voltage drops around the circuit is equal
to the impressed voltage source. Hence, we have
1
LI +
C ∫
I =0

or, in terms of the charge across the capacitor, we have the IVP
1
LQ + Q = 0 , Q (0) = 0 , Q (0) = 5 .
C
(c) The solution of the IVP is

sin ( 1
t )
Q (t ) = 5
LC
.
1
LC

This agrees with the oscillatory behavior predicted in part (a).

(d) With values L = 10 henries, C = 10−3 farads, the charge on the capacitor is

Q (t ) = 5
sin ( 100t ) = 1 sin10t .
100 2

 A Pendulum Experiment
59. The pendulum equation is
g
θ+ sin θ = 0 .
L

For small θ, we can approximate sin θ ≈ θ , giving the differential equation


g
θ + θ =0.
L
314 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 314

g
This is the equation of simple harmonic motion with circular frequency ω0 = , and natural
L

1 g 1 L
frequency f0 = . Hence, the period of motion is T = = 2π .
2π L f0 g

Tearth g sun
= = 400,000 = 100 40 ≈ 632 .
Tsun g earth

 Changing into Systems


60. 4 x − 2x + 3x = 17 − cost
x= y
1
y = (−3x + 2 y +17 − cost)
4
1
61. Lq + Rq + q = V (t)
c
q=I
1⎛ 1 ⎞
I= ⎜ − q − RI + V (t) ⎟
L⎝ c ⎠
1
62. 5q +15q + q = 5cos 3t
10
q=I
1
I =− q − 3I + cos3t
50

63. t 2 x + 4tx + x = t sin 2t t >0


4 1 sin 2t
x+ x+ 2 x=
t t t
x= y
x 4 sin 2t
y=− 2 − y+
t t t
64. 4 x +16x = 4sin t
x + 4x = sin t
x= y
y = −4x + sin t

 Circular Motion

65. Writing the motion in terms of polar coordinates r and θ and using the fact that the angular
velocity is constant, we have θ = ω0 (a constant). We also know the particle moves along a circle
of constant radius, which makes r a constant. We then have the relation x = r cosθ , and hence
315 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 315

x = ( −r sin θ )θ
x = ( −r sin θ )θ − ( r cos θ )θ 2 .

Because θ = 0 , θ = ω0 , we arrive at the differential equation

x + ω02 x = 0 .

 Another Harmonic Motion


66. For simple harmonic motion the circular frequency ω0 is
kR 2
ω0 = ,
mR 2 + I
so the natural frequency f 0 is
1 kR 2
f0 = .
2π mR 2 + I
 Motion of a Buoy

67. The buoy moves in simple harmonic motion, so the period is

2π = 2π m .
T = 2.7 =
ω0 k

We have one equation in two unknowns, but the buoyancy equation yields the second equation. If
we push the buoy down 1 foot, the force upwards will be F = V ρ , where V is the submerged
volume and ρ is the density of water. In this case, V = π r 2 h , r = 9 inches = 0.75 ft , h = 1 ft, and
9
ρ = 62.5 ft sec , so the force required to push the buoy down 1 foot is π (1)( 62.5 ) ≈ 110 lbs.
16

110
But k is the force divided by distance, so k = = 110 lbs ft . Finally, solving for m in the
1
kT 2

equation for T, we get m = , and substituting in all of our numbers, we arrive at m ≈ 20.4
4π 2

slugs (see Table 4.1.1. in the text.) The buoy weighs mg = ( 20.4 )( 32.2 ) = 657 lbs.

 Los Angeles to Tokyo

68. (a) Along the tunnel,

mx = −kr cosθ = −kx

x(0) = d if x is measured positive to the left of the center of the tunnel.


x(0) = 0 means that the train starts from rest (as soon as a brake is released).
316 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 316

(b) The solution to the IVP in part (a) is

x(t) = c1 cos ω0t + c2 sin ω0 t,


k
where ω0 = .
m

At the surface of the earth mg = kr

k q
where r = R, so ω0 = = .
m R

Letting x(0) = d yields c1 = d, while letting x(0) = 0 yields c2 = 0.

Hence we have
q
x(t) = d cos t.
R

For the train to go from L.A. to Tokyo, x(t) goes from d to −d.
q
and t goes from 0 to π.
R

Hence,

R
tf =π
q
4000 mi ×5280 ft/mi

32 ft/sec 2
= 2552 sec ≈ 42.5 minutes
R
(c) The solution tf = π from part (b) does not depend on the location of the points on the
q
earth’s surface; π, R, and q are all constant.

 Factoring Out Friction


69. (a) Letting x ( t ) = e( −b 2m )t X ( t ) , we have

−b ( −b 2m )t
X ( t ) + e(
−b 2m )t
x (t ) = e X (t )
2m
b 2 ( −b 2m )t −b ( −b 2m )t
x (t ) = 2
e X (t ) + e X ( t ) + e −( b 2m )t X ( t ) .
4m m
317 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.1 The Harmonic Oscillator 317

Substituting this into the original equation (1) and dividing through by e −( b 2m )t , we arrive
at
⎡ b b2 ⎤ ⎡ b ⎤
m⎢X − X + 2 X ⎥ + b ⎢−
X + X ⎥ + k[X] = 0.
⎣ m 4m ⎦ ⎣ 2m ⎦

Rearranging terms gives


⎡ b2 b2 ⎤
mX + [ −b + b ] X + − +k X =0
⎢ ⎥
⎣ 4m 2m ⎦
or
⎛ b2 ⎞
mX + ⎜ k − ⎟X =0.
⎝ 4m ⎠

b2
(b) If we assume k − > 0 , then divide by m and let
4m
1
ω0 = 4mk − b 2 )
2m

we find the solution of this DE in X is


X ( t ) = c1 cos ω0 t + c2 sin ω0t = Acos ( ω0 t − δ ) .
Thus, we have
− ( b 2m )t − ( b 2m )t
() () ( )
x t =e X t = Ae cos ω0t − δ .

 Suggested Journal Entry

70. Student Project


317 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 317

4.2 Real Characteristic Roots

 Real Characteristic Roots


1. y′′ = 0
The characteristic equation is r 2 = 0 , so there is double root at r = 0 . Thus, the general solution
is

y ( t ) = c e0t + c te0t = c + c t .
1 2 1 2

2. y′′ − y′ = 0

The characteristic equation is r 2 − r = 0 , which has roots 0, 1. Thus, the general solution is
y ( t ) = c1 + c2 et .

3. y′′ − 9 y = 0

The characteristic equation is r 2 − 9 = 0 , which has roots 3, –3. Thus, the general solution is
y ( t ) = c1e3t + c2e −3t .

4. y′′ − y = 0

The characteristic equation is r 2 −1 = 0 , which has roots 1, –1. Thus, the general solution is
y ( t ) = c1et + c2 e−t .

5. y′′ − 3y′ + 2 y = 0

The characteristic equation is r 2 − 3r + 2 = 0 , which factors into ( r − 2 )( r −1) = 0 , and hence has
roots 1, 2. Thus, the general solution is
y ( t ) = c1et + c2e 2t .

6. y′′ − y′ − 2 y = 0

The characteristic equation is r − r − 2 = 0 , which factors into ( r − 2 )( r +1) = 0 , and hence has
2

roots 2, –1. Thus, the general solution is


y ( t ) = c1e2t + c2 e −t .

7. y′′ + 2 y′ + y = 0

The characteristic equation is r 2 + 2r +1 = 0 , which factors into ( r +1)( r +1) = 0 , and hence has

the double root –1, –1. Thus, the general solution is


y ( t ) = c1e−t + c2 te−t .
318 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 318

8. 4 y′′ − 4 y′ + y = 0
The characteristic equation is 4r 2 − 4r +1 = 0 , which factors into ( 2r −1)( 2r −1) = 0 , and hence

1 1
has the double root , . Thus, the general solution is
2 2
y ( t ) = c1et 2 + c2tet 2 .

9. 2 y′′ − 3y′ + y = 0

The characteristic equation is 2r 2 − 3r +1 = 0 , which factors into ( 2r −1)( r −1) = 0 , and hence
1
has roots , 1. Thus, the general solution is
2
y ( t ) = c1et 2 + c2 et .

10. y′′ − 6 y′ + 9 y = 0

The characteristic equation is r 2 − 6r + 9 = 0 , which factors into ( r − 3)( r − 3) = 0 , and hence has
the double root 3, 3. Thus, the general solution is
y ( t ) = c1e3t + c2 te3t .

11. y′′ − 8y′ +16 y = 0


The characteristic equation is r 2 − 8r +16 = 0 , which factors into ( r − 4 )( r − 4 ) = 0 , and hence

has the double root 4, 4. Thus, the general solution is


y ( t ) = c1e4t + c2te 4t .

12. y′′ − y′ − 6 y = 0
The characteristic equation is r 2 − r − 6 = 0 , which factors into ( r + 2 )( r − 3) = 0 , and hence has

roots –2, 3. Thus, the general solution is


y ( t ) = c1e −2t + c2 e3t .

13. y′′ + 2 y′ − y = 0

(
The characteristic equation is r 2 + 2r −1 = 0 , which factors into r +1− 2 r +1 + 2 = 0 , and )( )
hence has roots −1+ 2 , −1− 2 . Thus, the general solution is

(
y ( t ) = e −t c e 2t
+ c e− 2t
).
1 2

14. 9 y′′ + 6 y′ + y = 0
The characteristic equation is 9r 2 + 6r +1 = 0 , which factors into ( 3r +1) = 0 , and hence has the
2

1 1
double root − , − . Thus, the general solution is
319 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 319
3 3
y ( t ) = c1e −t 3 + c2te −t 3 .
320 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 320

 Initial Values Specified


15. y′′ − 25y = 0 , y ( 0 ) = 1, y ′ ( 0 ) = 0

The characteristic equation of the differential equation is r 2 − 25 = 0 , which factors into


( r − 5)( r + 5) = 0 , and thus has roots 5, –5. Hence,
y ( t ) = c1e5t + c2 e −5t .

Substituting in the initial conditions y ( 0 ) = 1 gives c1 + c2 =1 . Substituting in y ( 0 ) = 0 gives


1
5c1 − 5c2 = 0 . Solving for c1 , c2 gives c1 = c2 = . Thus the general solution is
2
1 5t 1 −5t
y (t ) = e + e .
2 2
16. y′′ + y′ − 2 y = 0 , y ( 0 ) = 1, y ′ ( 0 ) = 0

The characteristic equation of the differential equation is r 2 + r − 2 = 0 , which factors into

( r + 2 )( r −1) = 0 , and thus has roots 1, –2. Thus, the general solution is
y ( t ) = c1e −2t + c2 et .
1 2
Substituting into y ( 0 ) = 1, y′ ( 0 ) = 0 yields c1 = , c2 = , so
3 3
1 2
y ( t ) = e −2t + et .
3 3
17. y′′ + 2 y′ + y = 0 , y ( 0 ) = 0 , y ′ ( 0 ) = 1

The characteristic equation is r 2 + 2r +1 = 0 , which factors into ( r +1)( r +1) = 0 , and hence has

the double root –1, –1. Thus, the general solution is


y ( t ) = c1e −t + c2 te .
−t

Substituting into y ( 0 ) = 0 , y ′ ( 0 ) = 1 yields c1 = 0 , c2 = 1, so

y ( t ) = te −t .
18. y′′ − 9 y = 0 , y ( 0 ) = −1, y ′ ( 0 ) = 0

The characteristic equation is r 2 − 9 = 0 , which factors into ( r − 3)( r + 3) = 0 , and hence has

roots are 3, –3. Thus, the general solution is


y ( t ) = c1e3t + c2 e −3t .
y ( 0 ) = −1,
Substituting into
321 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 321

y′ ( 0 ) = 0 yields c 1 2 =−
1
=c
, so
2
1 3t 1 −3t
y (t ) e − e .
=−
2 2
320 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 320

19. y′′ − 6 y′ + 9 y = 0 , y ( 0 ) = 0 , y ′ ( 0 ) = −1

The characteristic equation is r 2 − 6r + 9 = 0 , which factors into ( r − 3)( r − 3) = 0 , and hence has
the double root 3, 3. Thus, the general solution is
y ( t ) = c1e3t + c2 te3t .

Substituting into y ( 0 ) = 0 , y ′ ( 0 ) = −1 yields c1 = 0 , c2 = −1 , so

y ( t ) = −te3t .

20. y′′ + y′ − 6 y = 0 , y ( 0 ) = 1, y ′ ( 0 ) = 1

The characteristic equation is r 2 + r − 6 = 0 , which factors into ( r + 3)( r − 2 ) = 0 , and hence has

roots –3, 2. Thus, the general solution is


y ( t ) = c1e −3t + c2 e 2t .

1 4
Substituting into y ( 0 ) = 1, y′ ( 0 ) = 1 yields c1 = , c2 = , so
5 5
1 4
y ( t ) = e −3t + e 2t .
5 5
21. y′′ − y′ = 0 y(0) = 2, y(0) = −1

r2 − r = 0 (Characteristic equation)

r(r − 1) = 0 r = 0, 1

y = c1 + c2et ⇒ 2 = c1 + c2
y′ = c2et ⇒ −1 = c2, c1 = 3

y = 3 − et

22. y′′ − 4 y′ −12 y = 0 y(0) = 1, y′(0) = −1

r2 − 4r − 12 = 0 (Characteristic equation)

(r + 2)(r − 6) = 0 r = −2, 6
y = c e −2t + c e 6t y(0) = 1⇒ c + c = 1 7 ⎫ 1
1 2 1 2
⎬ ⇒ c = ,c =
−2t 6t ′ 1 2

y ′ = −2c1e + 6c2 e y (0) = −1⇒ −2c1 + 6c2 = −1⎭ 8 8

1 −2t 3 −6t
y= e + e
4 4
321 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 321

 Bases and Solution Spaces


23. y′′ − 4 y′ = 0
r2 − 4r = 0 (Characteristic equation)
r(r − 4) = 0 ⇒ r = 0, 4
Basis: {1, e4t}
Solution Space: {y ⎜y = c1 + c2e4t; c1, c2 ∈ \ }

24. y′′ −10 y′ + 25y = 0


r2 − 10r + 25 = 0 (Characteristic equation)
(r −5) = 0 ⇒ r = 5, 5
2

Basis: {e5t, te5t}


Solution Space: {y⎜y = c1e5t + c2te5t; c1, c2 ∈ \ }

25. 5y′′ −10 y′ −15 y = 0


5r2 − 10r − 15 = 0 (Characteristic equation)
r2 − 2r − 3 = 0 ⇒ (r − 3)(r + 1) = 0 ⇒ r = 3, −1
Basis: {e3t, e−t}
Solution Space: {y ⎜y = c1e3t + c2e−t; c1, c2 ∈ \ }
26. y′′ + 2 2 y′ + 2 = 0

r 2 + 2 2r + 2 = 0 (Characteristic equation)

(r + 2)(r + 2 ) = 0 r = − 2, − 2

Basis: e − { 2t
,te −
2t
}
Solution Space: {y y = c e− 2t
+ c te − 2t
;c ,c ∈ \}
1 2 1 2

 Other Bases
27. y′′ − 4 y = 0
r2 − 4 = 0 (Characteristic equation)
r = ±2 ∴ {e2t, e−2t} is a basis
To show {cosh 2t, sinh 2t} is a basis, we need only show that cosh 2t and sinh 2t are linearly
independent solutions:
cosh 2t sinh 2t
W= = 4 cosh2 2t − 4 sinh2 2t
2sinh 2t 2cosh 2t
322 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 322

2
⎛ e 2t + e −2t ⎞ ⎛ e 4t +1+ e −4t ⎞
⎟ =⎜
2
cosh 2t = ⎜ ⎟
⎝ 2 ⎠ ⎝ 2 ⎠
2
⎛ e 2t − e −2t ⎞ ⎛ e 4t −1− e −4t ⎞
⎟ =⎜
2
sinh 2t = ⎜ ⎟
⎝ 2 ⎠ ⎝ 2 ⎠
so cosh2 2t − sinh2 2t = 1 and W = 4 ≠ 0.
∴ cosh 2t, sinh 2t are linearly independent.
Substitute y = cosh 2t, y′ = 2 sinh 2t, y ′ = 4 cosh 2t

Then y′′ − 4y = 4 cosh 2t − 4 cosh 2t = 0 ∴ y = cosh 2t is a solution.


In similar fashion, we can show that y = sinh 2t is also a solution.

{
To show that e 2t ,cosh 2t } is a basis, we use the facts that e 2t
and cosh 2t are solutions.

Then:
e2t +e −2t
e 2t cosh 2t e 2t
W= = 2
2e 2t 2sinh 2t 2e 2t e 2t − e −2t

= (e4t − 1) − (e4t + 1) = −2 ≠ 0

∴ e2t and cosh 2t are linearly independent


28. y′′ = 0
r2 = 0 (Characteristic equation) so that r = 0, 0.
Basis: {1, t}
To show {t + 1, t − 1} is also a basis:
Note that for both y = t + 1 and y = t − 1, y ′ = 0, so both are solutions.
t +1 t −1
W= = (t + 1) − (t − 1) = 2
1 1
∴ t + 1, t − 1 are linearly independent
To show {2t, 3t − 1} is another basis:
Note that for both y = 2t and y = 3t − 1, y ′ = 0, so both are solutions.
2t 3t −1
W= - 6t − 2(3t − 1) = 2
2 3

∴ 2t, 3t − 1 are linearly independent


323 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 323

 The Wronskian Test


t +1 t −1 t 2 + t t3
1 2t + t 3t 2 t −1 t 2 + t t3
1 1 2t +1 3t 2
29. W= = (t +1) 0 2 6t −1 0 2 6t

0 0 2 6t
0 0 6 0 0 6
0 0 0 6

= (t + 1)12 − 12(t − 1) = 24 ≠ 0

Yes, {t + 1, t − 1, t2 + t, t3} is a basis for the solution space for y(4) = 0.


te5t e5t 2e5t −1 t 1 2 − e −5t

30. W= (5t +1)e5t 5e5t 10e5t = e5t 5t + 1 5 10


(25t +10)e5t 25e5t 50e5t 25t +10 25 50

⎡ 5 10 5t +1 10 5t +1 5 ⎤
= e5t t −1 + (2 − e −5t )

⎢ ⎥
⎣ 25 50 25t +10 50 25t +10 25 ⎦

= 25e5t ≠ 0
Yes, {te5t, e5t, 2e5t − 1} is a basis for the solution space for y′′′ −10 y′′ + 25y′ = 0.

31. The given set has only three solutions, so it cannot be a basis. A basis for the solution space of y(4)
= 0 must have 4 linearly independent solutions.

 Sorting Graphs

32.
324 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 324

 Relating Graphs

For Problems 33−35, x + 5x + 6x = 0 has (from Example 1) solutions


x(t) = c1e −2t + c2 e −3t (1)
x(t) = −2c1e −2t − 3c2e −3t (2)

33. (a), (b)

x(0) ≈ − 10 x(0) ≈ 0
⇓ ⇓
c1 + c2 = −10 −2c1 − 3c2 = 0

c1 = −30, c2 = 20
(c) From (1) in box, x(t) = −30e−2t + 20e−3t.

For t > 0, each term diminishes as t increases; the result remains negative, below the
t-axis.

For t < 0, each exponential increases as t decreases; the negative term cancels the positive
term when 30e−2t = 20e−3t or e−t = 1.5,
that is, when t = −ln 1.5 ≈ −.405 which looks about right on the tx-graph.
(d) From (2), x(t) = 60e−2t − 60e−3t = 60e−2t(1 − e−t) which is always positive for t > 0,

decreasing as t increases.

x(t) reaches a maximum when x (t) = −120e −2t +180e −3t = 0


−2 + 3e −t = 0
2
e −t = ,
3

2
so t = − ln ≈ 0.406, which looks about right on the tx -graph.
3
325 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 325

34. (a)

(b)
(b) x(0) ≈ 5 x(0) ≈ 0

⇓ ⇓
c1 + c2 = 5 −2c1 − 3c2 = 0
G G G G
Because all problems for finding ci are of type Ac = b , we solve for c = A −1b

⎡1 1⎤ ⎡ −3 −1⎤ G ⎡5⎤
−1
We have A = ⎢ ⎥ so A = − ⎢ ⎥ , and here b = ⎢ ⎥
⎣ −2 −3⎦ ⎣2 1⎦ ⎣0⎦

so
G ⎡−3 −1⎤ ⎡5⎤ ⎡+15⎤
c = −⎢ ⎥⎢ ⎥ = ⎢ ⎥
⎣ 2 1 ⎦ ⎣ 0 ⎦ ⎣ −10 ⎦

(c) From (1) in box on previous page, x(t)= 15e−2t − 10e−3t.

As t increases from zero, both exponentials decrease with their sum remaining positive, which
agrees with the tx graph.
−2t −3t −2t −t
(d) From (2), x(t) = −30e − 30e = −30e (1 + e )

For t > 0, this quantity is always negative, and as long as t increases, each term gets
closer to zero, in agreement with tx -graph.
326 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 326

35.

(a) x(0) ≈ 0 and x(0) ≈ −8

(b) From the method of 34(b),


⎡c1 ⎤ ⎡0⎤ ⎡−3 −1⎤ ⎡ 0 ⎤ ⎡ −8⎤
= A −1 =− =
⎢c ⎥ ⎢ −8⎥ ⎢2 1 ⎥⎦ ⎢⎣ −8⎦⎥ ⎢ +8 ⎥
⎣ 2⎦ ⎣ ⎦ ⎣ ⎣ ⎦

so from (1), x(t) = −8e−2t + 8e−3t.


−2t −3t
(c) For t > 0, e > e so the sum is always negative and approaches zero as t increases,

in agreement with the tx graph.


−2t −3t −2t −3t
(d) From (2) x(t) = +16e − 24e = 0, so 2e − 3e = 0,

2
which yields e−t = or t ≈ .406,
3

which looks about right on the tx -graph.


For t > 0.406, x(t) > 0 and decreases toward zero as t increases.
327 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 327

For Problems 36−39, x − x − 6x = 0 has (from Example 1) solutions

x(t) = c1e −2t + c2e3t (1)


−2t
x(t) = −2c1e + 3c2 e 3t
(2)

36. (a)

(b) From (1) c1 + c2 = 0


2 2
c1 = − ; c2 =
From (2) −2c1 + 3c2 = 2 5 5

2 2
x(t) = − e−2t + e3t
5 5
4 −2t 6 3t
x(t) = e + e
5 5

(c) For t > 0, e−2t < e3t, so x(t) is always positive, and as t increases, so does x(t).
This result agrees with the tx-graph.
For t < 0, e3t < e−2t so x(t) is always negative, and as t becomes more negative, x(t)
becomes more negative.
(d) x(t) is always positive.

For t > 0, e−2t < e3t, so the second term dominates as t increases, and x(t) increases as
well. These facts are in agreement with the tx -graph.
37. (a)
328 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 328

(b) From (1), c1 + c2 = 2 6 4


c1 = ; c =
From (2), −2c1 + 3c2 = 0 5 2 5

6 −2t 4 3t
x(t) = e + e
5 5
12 −2t 12 3t
x(t) e + e
=−
5 5

(c) x(t) is always positive.

For t > 0, as t increases, the first term decreases toward 0 and the second term increases
ever more rapidly, in agreement with the tx-graph.
(d) For t > 0, e3t > e−2t, so x(t) is positive, and x(t) increases as t increases, as shown on the
tx graph.

For t < 0, the first term will dominate and x(t) will be negative, ever more so as t

becomes more negative, in agreement with the tx -graph.

38. (a)

(b) From (1), c1 + c2 = −3 9 6


=−
From (2), −2c1 + 3c2 = 0 c1 = − ; c2
5 5
9 6
x(t) = − e −2t − e3t
5 5

18 18
x(t) e −2t − e3t
=+
5 5
(c) x(t) is always negative, with a maximum at t = 0. (See part (d) and set x(t) = 0.)
These facts agree with the tx graph.

(d) For t > 0, e3t > e−2t so the negative term dominates in x(t) and x(t) is negative, ever
more so as t increases.

For t < 0, e−2t > e3t so the positive term dominates in x(t) and x(t) is positive, ever more
329 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 329

so as t becomes more negative. These facts agree with the tx graph.


330 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 330

39. (a)

(b) From (1), c1 + c2 = 0 1 1


c1 = ; c2 = −
From (2), −2c1 + 3c2 = −1 5 5

1 −2t 1 3t
x(t) = e − e
5 5
2 −2t 3 3t
x(t) e − e
=−
5 5

(c) For t > 0 the second term dominates, so x(t) is negative, ever more so as t increases.

For t < 0 the first term dominates, so x(t) is positive, ever more so as t becomes more
negative.

These facts agree with the tx graph.


(d) x(t) is always negative. The maximum value will occur when t = 0, as shown on the tx -

graph.

 Phase Portraits

Careful inspection shows:

40. (B) 41. (D) 42. (A) 43. (C)


331 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 331

 Independent Solutions
44. Letting
r1t r2 t
c1e + c2 e =0

for all t, then by setting t = 0 and t = 1 we have, respectively


c1 + c2 = 0
c e r1 + c e r2 = 0.
1 2

When r1 ≠ r2 then these equations have the unique solution c1 = c2 = 0 , which shows the given

functions er1t , e r2t are linearly independent for r1 ≠ r2 .


 Second Solution

45. Substituting y = v ( t ) e −bt 2a into

ay ′ + by′ + cy = 0

gives
b −bt 2a
y′ = v′e −bt 2a − ve
2a
b b2
y ′′ = v′′e −bt 2a − v′e −bt 2a + 2 ve −bt 2a .
a 4a
Substituting v, v′, v ′ into the differential equation gives the new equation (after dividing by
e −bt 2a )

⎛ b b2 ⎞ ⎛ b ⎞
a ⎜ v′′ − v′ + 2 v ⎟ + b ⎜ v′ − v ⎟ + cv = 0 .
⎝ a 4a ⎠ ⎝ 2a ⎠

Simplifying gives
⎛ b2 ⎞
av′′ − ⎜ − c⎟v = 0 .
⎝ 4a ⎠
Because we have assumed b 2 = 4ac , we have the equation v ′ = 0 , which was the condition to be
proven.
 Independence Again
46. Setting
c1e −bt 2a + c2 te −bt 2a = 0

for all t, we set in particular t = 0 and then t = 1. These yield, respectively, the equations
c1 = 0
c1e −b 2a + c2 e −b 2a = 0

which have the unique solution c1 = c2 = 0 . Hence, the given functions are linearly independent.
331 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 331

 Repeated Roots, Long-Term Behavior


47. Because e −bt 2a approaches 0 as t → ∞ (for a, b > 0 ), we know the first term tends toward zero.

For the second term we need only verify that


t
te −bt 2a = bt 2a →0
e

does as well. To use l’Hôpital’s rule, we compute the derivatives of both the numerator and de-
nominator of the previous expression, getting
1
b bt 2a
,
2a
e

which clearly approaches 0 as t → ∞ . Then l’Hôpital’s rule assures us that the given expression
te −( b 2a )t approaches 0 as well.

 Negative Roots
We have r = −b ± b − 4mk , so in the overdamped case where b − 4mk > 0 , these
2 2
48.

characteristic roots are real. Because m and k are both nonnegative, b 2 − 4mk < b 2 causing

r1 = −b + b 2 − 4mk to be a negative sum of negative and positive terms

and
r2 = −b − b 2 − 4mk to be a negative sum of two negative terms.

 Circuits and Springs


1
49. (a) The LRC equation is LQ + RQ + Q = 0 , hence the following discriminant conditions
C
hold:

L
Δ = R2 − 4
C
<0 ( underdamped )
L
Δ = R2 − 4 = 0
C ( critically damped )
L
Δ = R2 − 4 > 0 ( overdamped ) .
C
332 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 332

(b) The conditions in part (a) can be written


L
R<2 ( underdamped )
C

R=2
L ( critically damped )
C

R>2
L ( overdamped ) .
C

1
These correspond to the analogy that m, b, and k correspond respectively to L, R, and .
C
(see Table 4.1.3 in the textbook.)
 A Test of Your Intuition
50. Intuitively, a curve whose rate of increase is proportional to its height will increase very rapidly
as the height increases. On the other hand, upward curvature doesn’t necessarily imply that the
function is increasing! (The curve e −t has upward curvature, yet decreases to 0 as t → ∞ .) In this

case, the restriction that y ′ ( 0 ) = 0 will cause the second curve to increase, but probably not

nearly as rapidly as the first curve. Solving the equations, the IVP y′ = y , y ( 0 ) = 1 has the

solution y = et , whereas the second curve described by y ′ = y , y ( 0 ) = 1, y ′ ( 0 ) = 0 has the


solution
1 t 1 −t
y (t ) = e + e .
2 2

The first curve is indeed above the second curve.

 An Overdamped Spring

51. (a) The solution of an overdamped equation has the form


x ( t ) = c1e r1t + c2e r2t .
Suppose that
c1 e r1t1 + c2 er2t1 = 0
rt rt ( r1 −r2 )t1
for some t1 . Because e 21
is never zero, we can divide by e 21
to get c1e + c2 = 0 .

Solving for t1 gives


1 −c
t1 = ln 2 .
r1 − r2 c1

This unique number is the only value for which the curve may pass through 0. If the
argument of the logarithm is negative or if the value of t1 does not
cross the
333 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 333

equilibrium point. is negative, then the solution

(b) By a similar argument, we can show that the derivative x ( t ) also has one zero.
333 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 333

 A Critically Damped Spring


52. (a) Suppose
( c1 + c2t1 ) er t
11
=0.

We can divide by the nonnegative quantity e r1t1 getting the equation c + c t = 0 , which
1 21
c
has the unique solution t1 = − 1 . Hence, the solution of a critically damped equation can
c2
pass through the equilibrium at most once. If the value of t1 is negative, then the solution
does not cross the equilibrium point.
(b) By a similar argument, we can show that the derivative x ( t ) has one zero.

 Linking Graphs
After inspection, we have labeled the yt and y′t graphs as follows.
y y' y'
53.
2 1 5 3 1 5
1 3
2
3
2
t t y
3 3 –5 5

4
4 4 t=0
–5 –5 –5

54.
334 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 334

55.

 Damped Vibration
56. The IVP problem is
1
x + 2x + x = 0 , x ( 0 ) = 3 in = ft , x ( 0 ) = 0 ft sec .
4

The solution is
1 −t 1 −t
x (t ) = e + te .
4 4
This is zero only for t1 = −1 , whereas the physical system does not start before t = 0 .

 Surge Functions

57. For mx + bx + kx = 0 , let m = 1, find b, k and initial conditions for the solution x = Ate−rt
x + bx + kx = 0
r 2 + br + k = 0 (characteristic equation)
−b ± b 2 − 4 ⋅1⋅ k
r=
2
b b
b2 − 4k = 0 to obtain repeated roots, r = − , −
2 2

x = c1e−rt + c2te−rt
x = −rc1e −rt + c2 (t(−r)e−rt + e −rt ) ∴ c1 = 0 = x(0)

c2 = A = x(0)

x(0) = c2

∴ b = −2r, 4k = b2

and from above we know 4k = 4r2 so that k = ±r, for k > 0


335 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 335

Results: r and A are given, and

b = −2r

k=r

x(0) = 0
x′(0) = A

 LRC-Circuit I
1
58. (a) LQ + RQ + Q = 2Q +101Q + 50Q = 0 , Q ( 0 ) = 99 , Q ( 0 ) = 0
C
(b) Q ( t ) = −e −50t +100e −t 2 (c) I ( t ) = Q ( t ) = 50e −50t − 50e −t 2

(d) As t → ∞ , Q ( t ) → 0 and I ( t ) → 0

 LRC-Circuit II
1
59. (a) LQ + RQ + Q = Q +15Q + 50Q = 0 , Q ( 0 ) = 5 , Q ( 0 ) = 0
C
(b) Q ( t ) = 10e −5t − 5e −10t (c) I ( t ) = Q ( t ) = −50e −5t + 50e−10t

(d) As t → ∞ , Q ( t ) → 0 and I ( t ) → 0

 The Euler-Cauchy Equation at 2 y′′ + bty′ + cy = 0


60. Let y ( t ) = t r , so

y′ = rt r −1
y ′′ = r ( r −1) t r −2 .

Hence
at 2 y ′′ + bty ′ + cy = ar ( r −1) t r + brt r + ct r = 0 .

Dividing by t r yields the characteristic equation


ar ( r −1) + br + c = 0 ,

which can be written as


ar 2 + ( b − a ) r + c = 0 .

If r1 and r2 are two distinct roots of this equation, we have solutions

y1 ( t ) = t r1
y2 ( t ) = t r2 .
336 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 336

Because these two functions are clearly linearly independent (one not a constant multiple of the
other) for r1 ≠ r2 , we have

y ( t ) = c1t r1 + c2 t r2

for t > 0 .

 The Euler-Cauchy Equation with Distinct Roots

For Problems 61–65, see Problem 60 for the form of the characteristic equation for the Euler-Cauchy DE.
61. t 2 y′′ + 2ty′ −12 y = 0

In this case a = 1 , b = 2 , c = −12 , so the characteristic equation is


r ( r −1) + 2r −12 = r 2 + r −12 = ( r + 4 )( r − 3) = 0 .

Hence, we have roots r1 = −4 , r2 = 3, and thus

y ( t ) = c1t 3 + c2 t −4 .

62. 4t 2 y′′ + 8ty′ − 3y = 0

In this case a = 4 , b = 8 , c = −3 , so the characteristic equation is


4r ( r −1) + 8r − 3 = 4r 2 + 4r − 3 = ( 2r −1)( 2r + 3) = 0 .
1 3
Hence, we have roots r1 = , r2 = − , and thus
2 2

y (t )= c1t
12 −3 2
+ c2t .

63. t 2 y′′ + 4ty′ + 2 y = 0

In this case a = 1 , b = 4 , c = 2 , so the characteristic equation is


r ( r −1) + 4r + 2 = r 2 + 3r + 2 = ( r +1)( r + 2 ) = 0 .

Hence, we have roots r1 = −1, r2 = −2 , and thus

y ( t ) = c1t −1 + c2 t −2 .
337 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 337

64. 2t 2 y′′ + 3ty′ − y = 0

In this case a = 2 , b = 3 , c = −1, so the characteristic equation is


2r ( r −1) + 3r −1 = 2r 2 + r −1 = ( 2r −1)( r +1) = 0 .

1
Hence, we have roots r1 = , r = −1, and thus
2 2
y ( t ) = c1t1 2 + c2 t −1 .

 Repeated Euler-Cauchy Roots

65. We are given that the characteristic equation


ar 2 + ( b − a ) r + c = 0
of Euler’s equation
at 2 y′′ + bty′ + cy = 0

has a double root of r. Hence, we have one solution y1 = t r . To verify that t r ln t is also a

r −1 r −1
solution, we differentiate y′ = rt ln t + t ,

y ′′ = r ( r −1) t r −2 ln t + rt r −2 + ( r −1) t r −2 = r ( r −1) t r −2 ln t + ( 2r −1) t r −2 .

By direct substitution we have


at 2 y ′′ + bty ′ + cy = at 2 ⎡⎣ r ( r −1) t r −2 ln t + ( 2r −1) t r −2 ⎤⎦ + bt ⎡⎣rt r −1 ln t + t r −1 ⎤⎦ + ct r ln t
= ⎡⎣ ar ( r −1) + br + c t r ln t + ⎡⎣ a ( 2r −1) + b t r .

We know that ar ( r −1) + br + c = 0 , so this last expression becomes simply

at 2 y ′′ + bty ′ + cy = ⎡⎣ a ( 2r −1) + b t r .

b −a
Thus the root of the characteristic equation is r = − , which makes this expression zero.
2a
b −a
To verify that t r and t r ln t are linearly independent (where r = − is the double
2a
root of the characteristic equation), we set
c1tr + c2t r ln t = 0

for specific values t = 1 and 2, which give, respectively, the equations


c1 = 0
338 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 338

c1 2 + c2 2 ln 2 = 0
r r

and yields the unique solution c1 = c2 = 0 . Hence, t and t ln t are linearly independent
r r

solutions.
339 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 339

 Solutions for Repeated Euler-Cauchy Roots


For Problems 66 and 67 use the result of Problem 60, y ( t ) = c1t r + c2 t r ln t .

66. t 2 y′′ + 5ty′ + 4 y = 0

In this case, a = 1 , b = 5 , and c = 4 , so our characteristic equation for r is r 2 + 4r + 4 = 0 , with a


double root at –2. The general solution is
y ( t ) = c1t −2 + c2t −2 ln t

for t > 0 .
67. t 2 y′′ − 3ty′ + 4 y = 0

In this case, a = 1 , b = −3 , and c = 4 , so our characteristic equation for r is r − 4r + 4 = 0 , with


2

a double root at 2. The general solution is


y ( t ) = c1t 2 + c2 t 2 ln t

for t > 0 .

68. 9t2y″ + 3ty′ + y = 0 Euler-Cauchy method: y = t m, t > 0


9m(m − 1) + 3m + 1 = 0 (characteristic equation)
9m2 − 6m + 1 = 0
1
(3m − 1)2 = 0 m=
3
y(t) = c1t1/ 3 + c2 t1/ 3 ln t

69. 4t 2 y′′ + 8ty′ + y = 0 Euler-Cauchy method: y = t m, t > 0


4m(m − 1) + 8m + 1 = 0
4m2 + 4m + 1 = 0
1
(2m + 1)2 = 0 m= −
2
y(t) = c1t −1/ 2 + c2 t −1/ 2 ln t

 Computer: Phase-Plane Trajectories


70. (a) y ( t ) = 2e−t + e −3t

The roots of the characteristic equation are –1 and –3, so the characteristic equation is

( r +1)( r + 3) = r 2 + 4r + 3 = 0 .
340 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 340

y ( t ) satisfies the differential equation

y ′ + 4 y′ + 3y = 0 .
341 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 341

(b) To find the IC for the trajectory of y ( t )


in yy′ space we differentiate y (t ) ,
getting

y ′ ( t ) = −2e −t − 3e −3t
The IC of the given trajectory of

( y ( t ) , y′ ( t ) )
in yy ′ space is ( y ( 0 ) , y′ ( 0 ) ) = ( 3, − 5) .

(c) We plot the trajectory starting at DE trajectories in yy ′ space

( 3, − 5) along with a few other

trajectories in yy′ space.

71. y ( t ) = e −t + e −8t

(a) The roots of the characteristic equation are –1 and –8, so the characteristic equation is

( r +1)( r + 8 ) = r 2 + 9r + 8 = 0 .
y ( t ) satisfies the differential equation

y ′ + 9 y′ + 8y = 0 .

(b) The derivative is


y ′ ( t ) = −e −t − 8e −8t .

The IC for the given trajectory in yy′


space is
( y ( 0 ) , y′ ( 0 ) ) = ( 2, − 9) .

(c) We plot this trajectory in yy′ space.

DE trajectory in yy′ space

72. y ( t ) = et + e−t

(a) The roots of the characteristic equation are 1 and –1, so the characteristic equation is

( r −1)( r +1) = r 2 −1 = 0 .
y ( t ) satisfies the differential equation

y′− y =0 .
342 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 342

(b) The derivative is y

y ′ ( t ) = et − e −t . 4

The IC for the given trajectory in yy′ 2


space is (2, 0)
y
( y ( 0 ) , y′ ( 0 ) ) = ( 2, 0) . –4 –2 2 4
–2
(c) We plot this and a few other trajectories
of this DE in yy′ space. –4

DE trajectories in yy′ space

73. y ( t ) = e −t + te −t

(a) The characteristic equation has a double root at –1, so the characteristic equation is

( r +1)
2
= r 2 + 2r + 1 = 0 .

y ( t ) satisfies the differential equation


y ′ + 2 y′ + y = 0 .

(b) The derivative is y'


1.5

y ′ ( t ) = −te −t .

The IC for the given trajectory in yy′ (1, 0) y

–1.5 1
1.5
space is
y ( 0 ) = 1, y ′ ( 0 ) = 0 .
–1.5

DE trajectory in yy′ space

74. y ( t ) = 3 + 2e 2t

(a) The roots of the characteristic equation are 0 and 2, so the characteristic equation is
r ( r − 2 ) = r 2 − 2r = 0 .

y ( t ) satisfies the differential equation

y ′ − 2 y′ = 0 .
341 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 341

(b) The derivative is


y ′ ( t ) = 4e 2t .

The IC for the given trajectory in yy′


space is

( y ( 0 ) , y′ ( 0 ) ) = ( 5, 4 ) .
(c) See the figure to the right.

DE trajectories in yy′ space


 Reduction of Order

75. (a) Let y2 = vy1 and


y2′ = v′y1 + vy1′
y2′ = v ′y1 + 2v′y1′ + vy1′.
Then
y2′′ + p ( x ) y2′ + q ( x ) y2 = v′′y1 + 2v′y1′ + pv′y1 + ( vy1′′ + pvy1′ + qvy1 ) = 0 .
Because y1′′ + py1′ + qy1 = 0 , cancel the terms involving v, and arrive at the new equation

y1v′′ + ( 2 y1′ + p ( x ) y1 ) v′ = 0

(b) Setting v′ = w and using the fact that y1′dx = dy1′ , we obtain

y1 w′ + ( 2 y1′ + p ( x ) y1 ) w = 0
⎛ 2y′1 +p (x )y1 ⎞
w′ + ⎜ ⎟w = 0
⎝ y1 ⎠
dw ⎛ −2 y1′−p (x )y1 ⎞
= ⎟ dx
w ⎜⎝ y1 ⎠
⎛ −2 y1′ ⎞
ln w = − p ( x ) dx

∫ ⎜⎝ y


−2
ln w = ∫y 1

dy1 − p ( x ) dx

ln w = ln y1−2 − p ( x ) dx

− p ( x ) dx
e ∫
w=± = v′
y12
− p ( x ) dx
e ∫
v=± ∫ y12
By convention, the positive sign is chosen.
342 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 342

(c) If v is a constant function on I, then v′ ≡ 0 and w ≡ 0 because v′ = w . The condition

w ≡ 0 contradicts our work in part (b) as ln w where w = 0 is undefined. Because v is

not constant on I, { y1 , y2 } is a linearly independent set of I.

 Reduction of Order: Second Solution


y′′ − 6 y′ + 9 y = 0 , y1 = e
3t
76.

We identify p ( t ) = −6 , so

∫ p ( t ) dt = −6t .
Substituting in the formula developed in Problem 75, we have
− p ( t ) dt
e ∫ e6t
y =y dt = e3t ∫ dt = te3t .

2 1 ∫ 2
3t 2

y1 ( t )
(e )
77. y′′ − 4 y′ + 4 y = 0 , y1 = e 2t

We won’t use the formula this time. We simply redo the steps in Problem 75. We seek a second
solution of the form y2 = vy1 = ve 2t . Differentiating, we have

y′2 = v′e 2t + 2ve 2t


y′′2 = v′′e 2t + 4v′e 2t + 4ve 2t .

Substituting into the equation we obtain


y2′′ − 4 y2′ + 4 y2 = v′′e2t = 0 .

Dividing by e2t gives v′′ = 0 or

v ( t ) = c1t + c2 .
Hence, we have found new solutions
y2 = ve 2t = c1te 2t + c2 e 2t .

Because y = e 2t , we let c =1, c = 0 , yielding a second independent solution


1 1 2

y2 = te2t .
343 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 343

78. t 2 y′′ − ty′ + y = 0 , y1 = t

We won’t use the formula this time. We simply redo the steps in Problem 75. We seek a second
solution of the form y2 = vy1 = tv . Differentiating, we have
y2′ = tv′ + v
y2′ = tv ′ + 2v′ .

Substituting into the equation we obtain


t 2 y2′ − ty′2 + y2 = t 3v ′ + t 2 v′ = 0 .

Letting w = v′ and dividing by t 3 yields


1
w′ + w = 0 .
t
We can solve by integrating the factor method, getting w = c1t −1 . Integrating we find
v = c1 ln t + c2 ,

so
y2 = tv = c1t ln t + c2t .

Letting c1 = 1, c2 = 0 , we get a second linearly independent solution

y2 = t ln t .

79. (t 2
)
+1 y ′′ − 2ty ′ + 2 y = 0 , y1 = t

We won’t use the formula this time. We simply redo the steps in Problem 75. We seek a second
solution of the form y2 = vy1 = tv . Differentiating yields
y2′ = tv′ + v
y2′ = tv ′ + 2v′ .

Substituting into the equation we get

(t 2
) ( )
+1 y2′ − 2ty2′ + 2 y2 = t t 2 +1 v ′ + 2v′ = 0 .

( )
Letting w = v′ and dividing by t t 2 +1 , we can solve the new equation using the integrating

factor method, getting


t2
∫(
2
dt = − ln t 2
+ 1 + (
2 ln t =)ln .
t t 2 +1 ) t 2 +1
344 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 344

We arrive at
t 2 +1
w=c = c + c t −2 .
1 1 1
t2
Integrating this, we get

(
v = c1 t − t −1 + c2 , )
so

(
y2 = tv = c1 t 2 −1 + c2 t . )
Letting c1 = 1, c2 = 0 we get a second linearly independent solution

y2 = t 2 −1.

 Classical Equations
80. y′′ − 2ty′ + 4 y = 0 , y1 ( t ) =1− 2t 2 (Hermite’s Equation)

( )
Letting y2 = vy1 = v 1− 2t 2 , we have

( )
y′2 = 1− 2t 2 v′ − 4tv

y ′′ = (1− 2t ) v′′ − 8tv′ − 4v


2
2

and perform the long division, yielding the equation


⎛ 8t ⎞
v ′ + ⎜ −2t + 2 ⎟ v′ = 0 .
⎝ 2t −1 ⎠

Letting w = v′ and solving the first-order equation in w, we get

( 2t )
−2
t2
w = c1e
2
−1 .

To find y2 we simply let c1 = 1 and integrate to get

∫ ( )
2 −2
v = et 2t 2 −1 dt .

( )
Multiplying by 1 − 2t 2 yields a final answer of

y2 ( t ) = (1− 2t ) ∫ e ( 2t −1)
2 t2 2 −2
dt .

81. (1− t ) y′′ − ty′ + y = 0 , y ( t ) = t


2
1 (Chebyshev’s Equation)

Letting y2 = vy1 = vt , we have y2′ = tv′ + v ,


345 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 345

y2′ = tv ′ + 2v′ ,
346 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 346

hence we have the equation


(1− t ) y ′ − ty′ + y
2
2 2 2
( 2
)
= t 1− t v ′ + 2 − 3t v′ = 0 . ( 2
)
( )
Dividing by t 1 − t 2 , and letting w = v′ ,

2 − 3t 2
w′ + w=0.
(
t 1− t2 )
Using partial fractions yields
2 − 3t 2 1 1
∫ t (1 − t ) dt = 2ln t + 2 ln t −1 + 2 ln 1+ t ,
2

so our integrating factor is t 2 1− t 2 and


1
w = c1 2
.
t 1− t 2
Letting c1 = 1 and multiplying by t yields a final answer of
1
y2 ( t ) = tv = t ∫t 2
1− t 2
dt .

This is a perfect example of a formula that does not tell us much about how the solutions behave.
Check out the IDE tool Chebyshev’s Equation to see the value of graphical solutions.
82. ty ′′ + (1− t ) y ′ + y = 0 , y1 ( t ) = t −1 (Laguerre’s Equation)

Letting y2 = vy1 = v ( t −1) , we have

y2′ = v′ ( t −1) + v , y2′ = v ′ ( t −1) + 2v′ ,

hence we have the equation

(
ty2′′ + (1− t ) y2′ + y = t ( t −1) v ′ + −t 2 + 4t −1 v′ = 0 . )
Dividing by t ( t −1) and letting w = v′ yields

−t 2 + 4t −1
w′ + w = 0.
t ( t −1)
1 2
Hence by use of partial fractions, our integrating factor is u = e ∫ t t −1 so that
−1+ + dt

t ( t − 1)
et
w = C1 .
t(t −1) 2
Letting c1 = 1 and multiplying by t −1 yields a final answer of

et
y2 ( t ) = v ( t −1) = ( t −1) ∫
347 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 347

2
dt .
348 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.2 Real Characteristic Roots 348

 Lagrange’s Adjoint Equation

83. (a)−(b) Differentiating the right side of


d
μ (t) [ y′′ + y ′ + y ] = [ μ (t) y′ + g(t) y ]
dt

we obtain
μ y′′ + μ y′ + μ y = μ ′y + μ y′ + g′y + gy′
Setting the coefficients of y, y′, y′′ equal, we find

for y : μ = μ (no information)


for y′ : μ = μ ′ + g

for y′′ : μ = g′

The last equation yields g = ∫ μdt and substituting this into the second equation, and
differentiating, gives a differential equation for the “integrating factor”
μ ′′ − μ ′ + μ = 0.

(c) We perform the differentiation on the right-hand-side of the given equation, yielding
μ (t) [ y′′ + p(t) y′ + q(t) y ] = μ y′′ + μ ′y′ + g(t) y′ + g′(t) y.

Multiplying out the left-hand side and subtracting yields


[ μ p(t) − μ ′ − g(t)] y′ + [ μ q(t) − g ′(t)] y = 0.
Setting the first set of coefficients equal to 0 yields μ ′ = μ p − g ,
hence μ ′′ = μ p′ + μ ′ p − g′, so that g′ = − μ ′′ + μ p′ + μ ′ p.

The second set of coefficients yields μ q − g′ = 0 so that g′ = μ q.

Setting these two equations for g′ equal to each other yields μ ′ − p μ ′ + (q − p′) μ = 0
which was to be shown.

 Suggested Journal Entry

84. Student Project


347 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 347

4.3 Complex Characteristic Roots

 Solutions in General
1. y′′ + 9 y = 0

The characteristic equation is r 2 + 9 = 0 , which has roots 3i, –3i. The general solution is
y ( t ) = c1 cos3t + c2 sin 3t .

2. y′′ + y′ + y = 0

1 3
The characteristic equation is r 2 + r +1 = 0 , which has roots − ± i . The general solution is
2 2
⎛ 3 3 ⎞
y ( t ) = e −t 2 c cos t + c sin t .
⎜ 1 2
2
2 ⎟
⎝ ⎠

3. y′′ − 4 y′ + 5y = 0

The characteristic equation is r 2 − 4r + 5 = 0 , which has roots 2 ± i . The general solution is


y ( t ) = e 2t ( c1 cost + c2 sin t ) .

4. y′′ + 2 y′ + 8 y = 0

The characteristic equation is r 2 + 2r + 8 = 0 , which has roots −1± i 7 . The general solution is

(
y ( t ) = e −t c1 cos 7t + c2 sin 7t . )
5. y′′ + 2 y′ + 4 y = 0

The characteristic equation is r 2 + 2r + 4 = 0 , which has roots −1± i 3 . The general solution is

(
y ( t ) = e −t c1 cos 3t + c2 sin 3t . )
6. y′′ − 4 y′ + 7 y = 0

The characteristic equation is r 2 − 4r + 7 = 0 , which has roots 2 ± i 3 . The general solution is

(
y ( t ) = e 2t c1 cos 3t + c2 sin 3t . )
7. y′′ −10 y′ + 26 y = 0

The characteristic equation is r 2 −10r + 26 = 0 , which has roots 5 + i . The general solution is
y ( t ) = e5t ( c1 cost + c2 sin t ) .
348 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 348

8. 3y′′ + 4 y′ + 9 y = 0

2 23
The characteristic equation is 3r 2 + 4r + 9 = 0 , which has roots − ± i . The general solution
3 3
is
⎛ 23 23 ⎞
y ( t ) = e −2t 3 c cos t + c sin t .
⎜ 1 3
2
3 ⎟
⎝ ⎠
9. y′′ − y′ + y = 0

1 3
The characteristic equation is r 2 − r +1 = 0 , which has roots ±i . The general solution is
2 2
⎛ 3 3 ⎞
y ( t ) = et 2 c cos t + c sin t .
⎜ 1 2
2
2 ⎟
⎝ ⎠
10. y′′ + y′ + 2 y = 0

1 7
The characteristic equation is r 2 + r + 2 = 0 , which has roots − ± i . The general solution is
2 2
⎛ 7 7 ⎞
y ( t ) = e −t 2 c cos t + c sin t .
⎜ 1 2
2
2 ⎟
⎝ ⎠

 Initial-Value Problems
11. y′′ + 4 y = 0 , y ( 0 ) = 1, y ′ ( 0 ) = −1

The characteristic equation is r 2 + 4 = 0 , which has roots ±2i . The general solution is
y ( t ) = c1 cos 2t + c2 sin 2t .

Substituting this into the initial conditions gives y ( 0 ) = c1 = 1, y ′ ( 0 ) = 2c2 = −1 . Hence, the
solution of the initial-value problem is

1
y ( t ) = cos 2t − sin 2t .
2
12. y′′ − 4 y′ +13y = 0 , y ( 0 ) = 1, y ′ ( 0 ) = 0

The characteristic equation is r 2 − 4r +13 = 0 , which has roots 2 ± 3i . The general solution is
y ( t ) = e 2t ( c1 cos 3t + c2 sin 3t ) .

Substituting this into the initial conditions yields y ( 0 ) = c1 = 1, y ′ ( 0 ) = 2c1 + 3c2 = 0 , resulting in
2
c1 = 1, c2 = − . Hence, the solution of the initial-value problem is
3
349 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 349

⎛ 2 ⎞
y ( t ) = e 2t ⎜ cos 3t − sin 3t ⎟ .
⎝ 3 ⎠
350 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 350

13. y′′ + 2 y′ + 2 y = 0 , y ( 0 ) = 1, y ′ ( 0 ) = 0

The characteristic equation is r 2 + 2r + 2 = 0 , which has roots −1± i . Hence, the general solution
is
y ( t ) = e −t ( c1 cos t + c2 sin t ) .

Substituting this into the initial conditions yields y ( 0 ) = c1 = 1, y ′ ( 0 ) = c1 − c2 = 0 , resulting in

c1 = 1, c2 = 1. Hence, the solution of the initial-value problem is

y ( t ) = e−t ( cos t + sin t ) .

14. y′′ − y′ + y = 0 , y ( 0 ) = 0 , y ′ ( 0 ) = 1

From Problem 6,
⎧ ⎡ 3 ⎡ 3 ⎫
y (t ) = e t 2
c cos (t ) + c sin (t ) .

⎪ ⎤ ⎤⎪
⎨ 1 ⎢ ⎥ 2 ⎢ ⎥⎬
2 2
⎪⎩ ⎢⎣ ⎥⎦ ⎢⎣ ⎥⎦ ⎪⎭

Substituting this into the initial conditions yields y ( 0 ) = 0 , y ′ ( 0 ) = 1 , resulting in c1 = 0 ,

2
c2 = 3 . Hence, the solution of the initial-value problem is
3
2 ⎛ 3 ⎞
y (t ) = 3e −t 2 sin

t .

3 ⎝ 2 ⎠
15. y′′ − 4 y′ + 7 y = 0 , y ( 0 ) = 0 , y ′ ( 0 ) = −1

From Problem 6,

y (t ) = e
2t
{c cos ( 3t ) + c sin ( 3t )} .
1 2

Subsituting this into the initial conditions yields y ( 0 ) = 0 , y ′ ( 0 ) = −1, resulting in

1
c1 = 0 , c2 = − 3.
3

Hence, the solution of the initial-value problem is

16.
y (t )
1
y′′ + 2 y′ + 5y = 0 ,
3e 2t sin ( 3t ) .
351 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 351

=−
3
y (0) y ′ ( 0 ) = −1
= 1,

The characteristic equation is r 2 + 2r + 5 = 0 , which has roots −1± 2i . Hence, the general solu-
tion is

y ( t ) = e −t ( c1 cos 2t + c2 sin 2t ) .
352 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 352

Subsituting this into the initial conditions yields y ( 0 ) = 1, y ′ ( 0 ) = −1, resulting in c1 = 1, c2 = 0 .

Hence, the solution of the initial-value problem is


y ( t ) = e −t cos 2t .

 Working Backwards
17. (r −1)3 = r 3 − 3r 2 + 3r −1
y′′′ − 3y′′ + 3y′ − y = 0

18. (r − 4)(r − (1− i))(r − (1+ i)) = r 3 − 6r 2 +10r − 8 = 0


y′′′ − 6 y ′ + 10 y′ − 8 y = 0

19. (r − 2)(r − (2 + i))(r − (2 − i)) = r 3 − 6r 2 + 13r −10


y′′′ − 6 y′′ +13y′ −10 y = 0

20. (r − 4)(r 2 − (2 + i))(r − (2 − i)) = r 4 − 4r 3 + r 2 +16r − 20 = 0


y (4) − 6 y′′′ + y′′ +16 y′ − 20 y = 0

 Matching Problems
21. y′′ − y′ = 0 ⇒ r = 0, 1
y(t) = c1 + c2et Graph D

22. y′′ + y′ = 0 ⇒ r = 0, −1
y(t) = c1 + c2e−t Graph B

23. y′′ + 3y′ + 2 y = 0 ⇒ r = −2, −1

y(t) = c1e−2t + c2e−t Graph A

24. y′′ − 5 y′ + 6 y = 0 ⇒ r = 2, 3

y(t) = c1e2t + c2e3t Graph C

−1 ± 3i
25. y′′ + y′ + y = 0 ⇒ r =
2
−⎛ ⎞
1

⎜⎝ 2 ⎠⎟ t ⎛ 3t 3t ⎞
y(t) = e c cos + c sin Graph G
⎜ 1
2
2
2 ⎟⎠

26. y′′ + y′ = 0 ⇒ r = ±i

y(t) = c1 cos t + c2 sin t Graph F

27. y′′ + 4 y′ + 4 y = 0 ⇒ r = −2,−2


353 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 353

−2t
y(t) = (c1 + c2 t)e Graph E
354 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 354

1 ± 3i
28. y′′ − y′ + y = 0 ⇒ r =
2

⎛1⎞
⎜ 2 ⎟t ⎛ 3 ⎞
y(t) = e ⎝ ⎠
c cos t + c sin 3 t Graph H
⎜ 1
2 2
2 ⎟
⎝ ⎠

 Euler’s Formula

29. (a) The Maclaurin series for e x is


1 2 1 3 1
e x = 1+ x + x + x +" + x n +"
2! 3! n!
1 1 1
eiθ = 1+ iθ + ( iθ ) ( iθ ) ( iθ )
2 3 n
(b) + +" + +"
2! 3! n!

(c) Using the given identities for i, we can write


1 1 1
eiθ = 1+ iθ + ( iθ ) + ( iθ ) + " + ( iθ ) + "
2 3 n

2! 3! n!
⎛ 1 1 ⎞ ⎛ 1 1 ⎞
= ⎜1 − θ 2 + θ 4 −" + ⎟ + i ⎜θ − θ 3 + θ 5 −" ⎟= cosθ + i sin θ
2! 4! 3! 5!
⎝ ⎠ ⎝ ⎠

(d) Done in part (c). (e) Done in part (c).

 Long-Term Behavior of Solutions


30. r1 < 0 , r2 < 0 . When r1 ≠ r2 , the solution is

y ( t ) = c1er1t + c2 e r2t

and goes to 0 as t → ∞ . When r = r1 = r2 < 0 , the solution has the form

y ( t ) = c1e rt + c2 te rt .

In this case using l’Hôpital’s rule we prove the second term te rt goes to zero as t → ∞ when
r<0.

31. r1 < 0 , r2 = 0 . The solution

y ( t ) = c1er1t + c2

approaches the constant c2 as t → ∞ because r1 < 0 .

32. r = α ± βi , y ( t ) = eα t ( c1 cos β t + c2 sin β t ) . For β ≠0 the solution y(t) oscillates with

decreasing amplitude when α < 0 ; oscillates with increasing amplitude when α > 0 ; oscillates

with constant amplitude when α = 0 .


355 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 355

33. r1 = 0 , r2 = 0 . The solution

y ( t ) = c1 + c2t

approaches ∞ as t → ∞ when c2 > 0 and –∞ when c2 < 0 .

34. r1 > 0 , r2 < 0 . The solution

y ( t ) = c1e r1t + c2 e r2t

approaches ∞ as t → ∞ when c1 > 0 and –∞ when c1 < 0 .



35. r = ± β i , y ( t ) = c1 cos β t + c2 sin β t is a periodic function of period , and amplitude
β
c12 + c22

 Linear Independence

36. Suppose
c1eα t cos β t + c2eα t sin β t = 0

on an arbitrary interval. Dividing both sides by eα t , then differentiating the new equation and
dividing by β, yields

c1 cos β t + c2 sin β t = 0
c2 cos β t − c1 sin β t = 0.

Hence, c1 = 0, c2 = 0 and we have proven linear independence of the given functions.

 Real Coefficients

37. Solution of the differential equation is


y ( t ) = k1eα t ( cos β t + i sin β t ) + k 2 eα t ( cos β t − i sin β t )
αt
= eα t ( k1 + k 2 ) cos β t + ie ( k1 − k2 ) sin β t .

For the solution to be real, there must exist real numbers r and s such that
k1 + k 2 = r
k1 − k 2 = si
Solving for k1 and k2 , we get
1 1
k1 = r + si
2 2
1 1
k 2 = r − si.
2 2
356 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 356

dny
 Solving =0
dt n

d4y
38. (a) =0 (b) y ( 4 ) = 0 . The characteristic equation is
dt 4
d3y r 4 = 0 , which has a fourth-order root at
= k3
dt 3 0. Hence, the solution is
d2y 2 3

= k 3t + k 2 y ( t ) = c0 + c1t + c2t + c3t ,


dt 2

dy 1 which is the same as in part (a).


= k t2 + k t + k
3 2 1
dt 2
1 3 1 2
y= k 3t + k2t + k1t + k 0
3! 2

(c) In general we have


1 1
= t n−1 + k t n−2 +" + k t + k
y ( t ) kn−1
( n −1)! n−2
( n − 2 )! 1 0

= cn−1t n−1 + cn−2 t n−2 +" + c1t + c0

because all of the constants are arbitrary.

 Higher-Order DEs
d5y 4d 4 y 4d 3 y
39. − + =0
5 4 3
dt dt dt

The characteristic equation is

( )
r 5 − 4r 4 + 4r 3 = r 3 r 2 − 4r + 4 = r 3 ( r − 2 ) = 0 ,
2

which has roots, 0, 0, 0, 2, 2. Hence,


y ( t ) = c + c t + c t + c e + c te .
2 2t 2t

1 2 3 4 5

d 3 y 4d 2 y 7dy
40. + − −10 y = 0
dt 3 dt 2 dt

The characteristic equation is

r 3 + 4r 2 − 7r −10 = 0 ,
which has roots, –1, 2, –5. Hence,
357 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 357

y ( t ) = c e−t + c e 2t + c e −5t . 1 2 3
358 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 358

d 5 y dy
41. − =0
dt 5 dt

The characteristic equation is

( ) ( )( )
r 5 − r = r r 4 −1 = r r 2 −1 r 2 +1 = r ( r −1)( r +1) r 2 + 1 = 0 , ( )
which has roots, 0, ±1, ±i. Hence
y ( t ) = c1 + c2 et + c3e −t + c4 cost + c5 sin t .

42. y′′′ − 4 y ′ + 5y′ − 2 y = 0


r 3 − 4r 2 + 5r − 2 = 0 (characteristic equation)
f (1) = 1− 4 + 5 − 2 = 0 By long division, we obtain
∴r = 1 is a root (r −1)(r 2 − 3r + 2) = 0

(r −1)(r − 2)(r −1) = 0 r = 1, 1, 2

y(t) = c1et + c2 tet + c3e 2t


43. y′′′ + 6 y′′ +12 y′ + 8y = 0
r 3 + 6r 2 +12r + 8 = 0 (characteristic equation)

f (−2) = −8 + 24 − 24 + 8 = 0 By long division, we obtain


∴r = −2 is a root (r + 2)(r 2 + 4r + 4) = 0
(r + 2)3 = 0 r = −2, −2, −2

y(t) = c1e −2t + c2 te−2t + c3t 2 e −2t


44. y (4) − y = 0
r 4 −1 = 0 (characteristic equation)

(r 2 +1)(r 2 −1) = 0 r = ±i,±1

y(t) = c1 cost + c2 sin t + c3e t + c4 e −t

 Linking Graphs
y y' y'
45. 3
5 5 5
3

1
1
t t y
3 1 –5 1 5

2 3
–5 –5 –5 t=0
359 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 359

46. y y' y'


5 3 5 5 3

1 2
2
t t y
3 1 –5 5
2
1 1
3
–5 –5 –5 t=0

 Changing the Damping


47. The curves below show the solution of
x + bx + x = 0 , x ( 0 ) = 4 , x ( 0 ) = 0

for damping b = 0 , 0.5, 1, 2, 4. The larger the damping the faster the curves approach 0. The

curve that oscillates has no damping ( b = 0 ) .


.
x
x(t) b = 4
b=1 4
b=0
4 b=2
2
2 b = 0.5 b=2
b=4
t x
4 8 12 16 –4 –2 2 4
–2
–2
b = 0.5
–4 b=1
b=0
–4

In Figure 4.3.12 (b) in the text the larger the damping b the more directly the trajectory “heads”
for the origin. The trajectory that forms a circle corresponds to zero damping. Note that every
time a curve in (a) crosses the axis twice the corresponding trajectory in (b) circles the origin.
360 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 360

 Changing the Spring

48. (a) The solutions of x(t)


k =1
4
x + x + kx = 0 , x ( 0 ) = 4 , x ( 0 ) = 0 k = 0.5
2

are shown for k =


1
,
1
, 1, 2, 4. For k = 0.25

4 2 t
4 8 12 16
larger k we have more oscillations. k =2
–2 k =4

–4
.
(b) For larger k, since there are more oscilla- x

tions, the phase-plane trajectory spirals


4
further around the origin.
k =2 2
k = 0.25
–4 –2 2 x
–2
–4 k = 0.5

k =4 k =1
 Changing the Mass
k
49. (a) b = 0 and ωo = so that ωo is inversely proportional to m.
m
1
(b) If m is doubled, ωo is decreased by a factor of .
2
(c) If m is doubled, the damping required for critical damping is increased by a factor of 2.
 Finding the Maximum
50. (a) x + 2x + 3x = 0 , x(0) = 1, x(0) = 0

r2 + 2r + 3 = 0 (characteristic equation)
r = −1± 2i

(
x = e −t c1 cos 2t + c2 sin 2t ) ⎫

⎬ ⇒

( ) (
x = e −t − 2c1 sin 2t + 2c2 cos 2t − e −t c1 cos 2t + c2 sin 2t ) ⎪

1 = c1
1
0= 2c2 −1 so that c2 =
2
361 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 361

⎛ 1 ⎞
x = e−t ⎜ cos 2t + sin 2t ⎟
⎝ 2 ⎠
362 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 362

To find maximum displacement, set x = 0 .

⎛ ⎛ 1 ⎞ ⎞ ⎛ 1 ⎞
x = e −t − 2 sin 2t + 2 cos 2t − e −t cos 2t + sin 2t = 0
⎜ ⎜ ⎟ ⎟ ⎜ ⎟
⎝ ⎝ 2⎠ ⎠ ⎝ 2 ⎠
⎛ 2 ⎞ π
− 2+ sin 2t = 0 when 2t = π , so that t = sec
⎜ 2 ⎟ 2
⎝ ⎠
π π

− ⎛ ⎛ π ⎞ 1 π ⎞
⎟+ ⎟ = −e
2 2
Substituting for t: xmax = e ⎜ cos 2 ⎜ sin 2
⎝ ⎝ 2⎠ 2 2⎠
π

2
Max. Amplitude xmax = e

(b) m = 1, b = 2, k = 10 x(0) = 0, x(0) = 2

The DE is x′′ + 2x′ +10x = 0 for which the characteristic equation

r2 + 2r + 10 = 0 gives r = −1 ± 3i.
x(t) = e−t (c1 cos 3t + c2 sin 3t)
x′(t) = e−t (−3c1 sin 3t + 3c2 cos 3t) − e −t (c1 cos 3t + c2 sin 3t)

2 −t
x(t) = e sin 3t is the solution
3

To find the maximum displacement x max , set x′(t) =0 and solve for t:

2
0 = (3e−t cos 3t − e− t sin 3t)
3
so that tan 3t = 3 and t = 0.416 radians which gives x max = 0.4172 .

(c) m = 1, b = 4, k = 4 x(0) = 0, x(0) = 2

The DE is x′′ + 4x′ + 4 = 0 for which the characteristic equation

r2 + 4r + 4 = 0 gives r = −2, −2
x(t) = c1e −2t + c2te −2t ⎫⎪
⎬ ⇒ c1 = 0, c2 = 2
x′(t) = −2c1e −2t + c2 (−2te −2t −2t
+ e ) ⎪⎭
−2t
The solution is x(t) = 2te .

To find the maximum displacement xmax, we set so that t = 1/2 which


0 = 2(−2te −2t + e −2t ) yields xmax = e−1.
363 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 363

x′(t) = 0 and solve for t:


358 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 358

 Oscillating Euler-Cauchy
51. We used the substitution y = t r and obtained for r1 = α + i β and r2 = α − i β the solution

y ( t ) = k1t α +iβ + k2 t α −iβ = k1 e(α +iβ ) ln t + k2 e(α −iβ ) ln t = k1 eα ln t +iβ ln t + k2 eα ln t −iβ ln t

=e
α ln t
( c1 cos ( β ln t ) + c2 sin ( β ln t )) = tα ( c1 cos ( β ln t ) + c2 sin ( β ln t ) ).

This is the same process as that used at the start of Case 3 in the text utilizing the Euler’s For-
mula (4).
1 3
52. t 2 y′′ + 2ty ′ + y = 0 , r ( r −1) + 2r +1 = 0 , r 2 + r +1 = 0 , r = − ± i,
2 2
⎡ ⎛ 3 ⎞ ⎛ 3 ⎞⎤
y (t ) = t −1 2
c cos ln t + c sin ln t

⎢ 1 ⎜ 2 ⎟ 2 ⎜ 2 ⎟⎥
⎝ ⎠ ⎝ ⎠
53. t 2 y′′ + 3ty′ + 5 y = 0

Letting y = t r yields
t 2 r ( r −1) t r −2 + 3trt r −1 + 5t r = 0
t r {r ( r −1) + 3r + 5} = 0,

and gives r 2 + 2r + 5 = 0 , which has roots −1± 2i . Hence, the solution is

y (t ) = t
−1
( c1 cos ( 2 ln t ) + c2 sin ( 2 ln t ) ) .

54. t 2 y′′ + 17ty′ +16 y = 0 Euler-Cauchy: y = tm, t > 0

m(m − 1) + 17m + 16 = 0 (characteristic equation)


m2 + 16m + 16 = 0
−16 ± (16) 2 − 4(16)
m= = −8 ± 4 3
2

(
y(t) = t −8 c1 cos(4 3 ln t) + c2 sin(4 3 ln t) )
359 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 359

 Third-Order Euler-Cauchy
55. The third-order Euler-Cauchy equation has the form at 3 y′′′ + bt 2 y ′ + cty′ + dy = 0 . The derivatives
of y = t r (t > 0) are

y′ = rt r −1
y ′′ = r ( r −1) t r −2
y ′′′ = r ( r −1)( r − 2 ) t r −3

Substitute these equations into the third-order Euler-Cauchy equation above to obtain:
at 3 r ( r −1)( r − 2 ) t r −3 + bt 2 r ( r −1) t r −2 + ctrt r −1 + dt r = 0
at r r ( r −1)( r − 2 ) + bt r r ( r −1) + ct r r + dt r = 0

Dividing by t r , we obtain the characteristic equation:


ar ( r −1)( r − 2 ) + br ( r −1) + cr + d = 0

 Third-Order Euler-Cauchy Problems


56. t 3 y′′′ + t 2 y′′ − 2ty′ + 2 y = 0 has Euler-Cauchy characteristic equation:

r ( r −1)( r − 2 ) + r ( r −1) − 2r + 2 = 0
r 3 − 3r 2 + 2r + r 2 − r − 2r + 2 = 0
r 3 − 2r 2 − r + 2 = 0
Note: r = 1 is a zero of the polynomial f ( r ) = r 3 − 2r 2 − r + 2 because

f (1) = 1− 2 −1+ 2 = 0 .

Therefore r −1 is a factor of r 3 − 2r 2 − r + 2 , which enables us to find the other factors.


r 3 − 2r 2 − r + 2 = ( r −1)( r +1)( r − 2 )

so r = 1, –1, 2. Hence, the general solution to this Euler-Cauchy DE is


y ( t ) = c1t + c2t −1 + c3t 2 ,

for t > 0 .
57. t 3 y′′′ + 3t 2 y′′ + 5ty = 0 Let y = tm, t > 0

m(m −1)(m − 2) + 3m(m −1) + 5m = 0 (characertistic equation)


m − 3m + 2m + 3m − 3m + 5 = 0
3 2 2

m3 + 4m = 0 m = 0, ±2i

y(t) = c1 + c2 cos (2 ln t) + c3 sin (2 ln t)


360 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 360

 Inverted Pendulum
58. The differential equation x − x = 0 has the characteristic equation r 2 −1 = 0 with roots ±1.

Hence, the general solution is


x ( t ) = c1et + c2 e−t .

1 1
(a) With initial conditions x ( 0 ) = 0 , x ( 0 ) = 1 , we find c1 = and c2 = − . Hence, the solu-
2 2
tion of the IVP is
1 t 1 −t
x (t ) = e − e = sinh t .
2 2

(b) As t → ∞ , x ( t ) → 0 if c1 = 0 , and then x ( t ) → 0 also. This will happen whenever

x ( 0 ) = −x ( 0 ) .
 Pendulum and Inverted Pendulum
59. (a) The inverted pendulum equation has characteristic equation r 2 −1 = 0 , which has roots
±1. Hence, the solution

x ( t ) = c1et + c2 e = c1 ( cosh t + sinh t ) + c2 ( cosh t − sinh t ) = C1 sinh t + C2 cosh t ,


−t

where C1 = c1 − c2 , C2 = c1 + c2 .

(b) The characteristic equation of the pendulum equation is r 2 +1 = 0 , which has roots ±i .
Hence, the solution

x ( t ) = c1 cost + c2 sin t .

(c) The reader may think something strange about this because one form (a) appears real and
(b) complex, but they are really the same; the difference is taken up by how one chooses
{ }
the coefficients c1 , c2 in each case. The span of eit , e −it is the same as the span of
{sin t, cos t} .
361 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 361

 Finding the Damped Oscillation


60. The initial conditions x
1.5
x (0) = 1 , x (0) = 1 x(t) = e−t (cost + 2 sin t)
1
give the constants c1 = 1, c2 = 2 . Hence, we have

x ( t ) = e −t ( cost + 2sin t ) . 0.5

t
1 3 5

–0.5

 Extremes of Damped Oscillations


61. The local maxima and minima of the curve
x ( t ) = eα t ( c1 cos ω t + c2 sin ω t )

have nothing to do with the exponential factor eα t ; they depend only on


c1 cos ω t + c2 sin ω t ,

which can be rewritten as Acos (ω t − δ ) having period T = . Hence, consecutive maxima and
ω
minima occur at equidistant values of t, the distance between them being one-half the period, or
π
. (You can note in Problem 32 that the time between the first local maxima and the first local
ω
π
minima is = π .)
1

 Underdamped Mass-Spring System


62. We are given parameters and initial conditions
m = 0.25 , b = 1, k = 4 , x ( 0 ) = 1 , x ( 0 ) = 0 .

Hence, the IVP is


0.25 x + x + 4x = 0 , x ( 0 ) = 1 , x ( 0 ) = 0 ,

which has the solution


⎛ 3 ⎞
x ( t ) = e −2t cos 2 3t + sin 2 3t .
⎜ ⎟
⎝ 3 ⎠
362 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 362

 Damped Mass-Spring System


63. The IVP is
x + bx + 64x = 0 , x ( 0 ) = 1 , x ( 0 ) = 0 .

⎛ 5 ⎞
(a) b = 10 : (underdamped), x ( t ) = e −5t cos 39t + sin 39t
⎜ ⎟
⎝ 39 ⎠

b = 16 : (critically damped), x ( t ) = (1+ 8t ) e


−8t
(b)

(c) b = 20 : (overdamped), x ( t ) =
1
3
(4e −4t − e −16t )

 LRC-Circuit I
1
64. (a) The IVP is LQ + RQ + Q = Q + 8Q + 25Q = 0 , Q ( 0 ) = 1 , Q ( 0 ) = 0
C
⎛ 4 ⎞ 5 4
(b) Q ( t ) = e −4t cos3t + sin 3t = e −4t cos ( 3t − δ ) where δ = tan −1
⎜ 3 ⎟ 3 3
⎝ ⎠
20 4
(c) I ( t ) = Q ( t ) = −5e −4t sin ( 3t − δ ) − e −4t cos ( 3t − δ ) where δ = tan −1
3 3

(d) Charge on the capacitor and current in the circuit approach zero as t → +∞ .

 LRC-Circuit II
1 1
65. (a) The IVP is LQ + RQ + Q = Q + 1Q + 4Q = 0 , Q ( 0 ) = 1 , Q ( 0 ) = 0
C 4
⎛ ⎞ 2 3 −2t
(b) Q ( t ) = e −2t cos 2 3t +

3
sin 2 3t =

e cos 2 3t − δ , tan δ = ( 3
)
⎝ 3 ⎠ 3 3

(c) I (t ) = Q (t ) = −
4 3 −2t
( ) (
e cos 2 3t − δ − 4e −2t sin 2 3t − δ , tan δ =
3
)
3 3
(d) As t → ∞ , both Q ( t ) → 0 and I ( t ) → 0

 Computer Lab: Damped Free Vibrations

66. IDE Lab


363 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 363

 Effects of Nonconstant Coefficients


1
67. x+ x=0
t

(a) This ODE describes (among other things) an undamped vibrating spring in which the
restoring force is initially very large (when t is near zero), but eventually decays to zero,
causing the frequency of vibration to decrease and the solution period to increase as t
increases.
(b) We plotted the solution with IC x ( 0.1) = 2 , x ( 0.1) = 0 in the tx and xx planes.

x
3
2
1

–10 x
–20 10 20
–1
–2
–3
–4

(c) As we expected, the tx graph shows that the period of the oscillation increases with t. We
see also that the amplitude increases in the absence of friction. The xx phase portrait
shows that as time and amplitude increase, velocity decreases, which is consistent with
the previous observations. A good question for further exploration would be whether
amplitude increases indefinitely or levels off.
1
68. x+ x+x=0
t

(a) This ODE describes a damped vibrating spring in which the damping starts very large
when t is near zero, but decays to zero. We suspect that initially the amplitude of a
solution will rapidly decay, but as time increases the motion could become almost like
simple harmonic oscillation, as there will be almost no friction.
(b) We plotted the solution with IC x ( 0.1) = 2 , x ( 0.1) = 0 in the tx as well as the xx
planes.
364 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 364

x
1
0.5

–1 1 2 x
–0.5
–1
–1.5
–2

(c) As first expected, the tx graph shows that the solution is rapidly decaying. However the
xx phase portrait, constructed with a longer time interval, shows that our second expectation
is not confirmed. As time increases the oscillations do not become harmonic—the
amplitude of the oscillations continues to decrease, gradually and indefinitely.

tx + x = 0
69.

(a) If you divide by t, you will see that this equation is the same as the equation in
Problem 67.

70. ( )
x + x 2 −1 x + x = 0

(a) This ODE shows negative friction for x < 1 and positive damping for x > 1. For a small

initial condition near x = 0 , we might expect the solution to grow and then oscillate

around x = 1.

(b) We plotted the solutions in the tx and xx planes at initial velocity x ( 0 ) = 0 for three

different initial displacements: x ( 0 ) = 0.5 , x ( 0 ) = 2.0 , x ( 0 ) = 4.0 .


x.
4

–4 4 x

–4
365 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 365

(c) As expected, the tx graph shows that initially the solution is growing for x ( 0 ) = 0.5 and

decaying for x ( 0 ) = 4 . We also see that all the solutions seem to become periodic with

the same amplitude and period, but we note that the motion is not exactly sinusoidal and
that the amplitude is about 2 rather than 1 as we suspected. The xx phase portrait
confirms that the long term trajectories are not circular as in simple harmonic motion, but
distorted as we see in the tx graph. This equation is called van der Pol’s equation and
describes oscillations (mostly electrical) where internal friction depends on the value of the
dependent variable x; further details will be explored in Chapter 7.
71. x + ( sin t ) x + x = 0

(a) In this ODE damping changes periodically from negative to positive, so we can predict
oscillation in amplitude as well as periodic vibratory motion.
(b) We plotted the solution with IC x ( 0 ) = 2, x ( 0 ) = 0 in the tx and xx planes.

(c) The tx graph looks like a superposition of two periodic oscillations. The xx phase
portrait for a longer time interval shows that continued oscillations almost repeat, but never
exactly. This is called quasi-periodic motion.
1
72. x + x + tx = 0
t

(a) For this ODE damping is initially large, but vanishes as time increases; the restoring
force on the other hand is initially small but increases with time. How will these effects
combine?
(b) We plotted the solution with IC x ( 0.1) = 2, x ( 0.1) = 0 in the tx and xx planes.
366 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 366

x
2

–1 1 2 x

–1

–2
(c) As we expected, the tx graph shows initially large damping, which rapidly decreases the
amplitude of the solution, and increasing frequency, due to the effect of the increasing
spring “constant”, which shortens the period. The center of the xx graph will continue to
fill in, very slowly, if you give it a much longer time interval.
73. x + ( sin 2t ) x = 0

(a) In this ODE the restoring force changes periodically from positive to negative with a
frequency that is different from the natural frequency of the spring. We expect some
complicated but periodic motion.
(b) We plotted the solution with IC x ( 0 ) = 2 , x ( 0 ) = 0 in the tx and xx planes.

20 40 60 80 100 t
–2

–4

(c) The tx graph to t = 100 indeed looks almost periodic, with period 50. However the xx

phase portrait over a longer time interval shows that continued motion almost repeats, but
never exactly. This is another example of quasi-periodic motion, as in Problem 71.
Extending the tx graph will be another good way to see that the long term motion is
indeed not perfectly repeating.
367 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 367

 Boundary-Value Problems
⎛π⎞
74. y′′ + y = 0 , y(0) = 0, y ⎜ ⎟ = 0
⎝ 2⎠

y(t) = c1 cos t + c2 sin t


y(0) = 0 = c1
⎛π⎞
y = 0 = c , so y(t) = 0 is the solution.
⎜2⎟ 2
⎝ ⎠
⎛π⎞
75. y′′ + y = 0 , y(0) = 0, y ⎜ ⎟ = 1
⎝ 2⎠

y(t) = c1 cos t + c2 sin t


y(0) = 0 = c1
⎛π⎞
y = 1 = c , so y(t) = sin t is the solution.
⎜2⎟ 2
⎝ ⎠
76. y′′ + y = 0 , y(0) = 1, y(π ) = 1

y(t) = c1 cos t + c2 sin t

y(0) = 1 = c1
y(π ) = 1 = −c1 *No solutions
⎛π⎞ ⎛π⎞
77. y′′ + y = 0 , y ⎜ ⎟ = 1, y ⎜ ⎟ = 2
⎝4⎠ ⎝2⎠

y(t) = c1 cos t + c2 sin t


1 1
1 = c1 + c2 c1 + c2 = 2
2 2

2 = c2 c1 = 2 −2

y(t) = ( )
2 − 2 cost + 2sin t is the solution.
368 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 368

 Exact Second-Order Differential Equations

1 1 ⎡ 1 ⎤′
78. y ′′ + y′ − 2 y = 0 is the same as y′′ + ⎢ y ⎥ = 0 .
t t ⎣t ⎦
1
Integrating we obtain the linear equation y′ + y = c1 ,
t
1
dt

for which μ = e
∫t
= e ln t = t so we have ty′ + y = c t.
1

d c c c
Thus, (ty) = c1t, so ty = 1 t 2 + c2 and y(t) = 1 t + 2 .
dt 2 2 t
c
Substituting back into the original equation we find c1 = 0, so y(t) = is the general solution.
t
2 2
79. y′′ + y′ − 2 = 0
t t
⎡ 2 ⎤′
y′′ + y =0
⎢⎣ t ⎦⎥

Integrating and setting c1 = 0, we obtain


2

2
y + y=0 ∫ dt
μ = e t = e 2ln t = t 2
t

t 2 y′ + 2ty = 0 = c.
d 2 c
(t y) = 0 , t2y = c, y(t) = 2
dt t

80. (t 2 − 2t) y′′ + 4(t −1) y′ + 2 y = 0 where t 2 − 2t ≠ 0

Find (gy)′′ :
(gy)′ = gy′ + yg′
(gy) ′ = (gy′ + yg′)′ = gy ′ + y′g′ + yg ′ + g′y′
= gy′′ + 2g′y′ + yg′′
Let g = t2 − 2t. Then g′ = 2t − 2, g ′ = 2

Then (gy)′′ = (t 2 − 2t) y′′ + 4(t −1) y′ + 2 y


(gy)′′ = 0
(gy)′ = c1
gy = c1t + c2
c1t+c2
so y(t) = .
t 2 − 2t
369 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.3 Complex Characteristic Roots 369

 Suggested Journal Entry

81. Student Project


369 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 369

4.4 Undetermined Coefficients

 Inspection First
1. y′′ − y = t ⇒ y p ( t ) = −t 2. y′′ + y ′ = 2 ⇒ y p ( t ) = 2t

3. y′′ = 2 ⇒ y p ( t ) = t 2 4. ty′′ + y ′ = 4t ⇒ y p ( t ) = t 2

5. y′′ − 2 y′ + 2 y = 4 ⇒ y p ( t ) = 2 6. y′′ − y = −2cost ⇒ y p ( t ) = cost

7. y′′ − y′ + y = et ⇒ y p ( t ) = et 8. y ′′′ + y ′ + y = 2t + 2 ⇒ y p ( t ) = 2t

 Educated Prediction
The homogeneous equation y′′ + 2 y′ + 5y = 0 has characteristic equation r 2 + 2r + 5 = 0 , which has
complex roots −1± 2i . Hence,

yh ( t ) = c1e−t sin 2t + c2 e−t cos 2t ,

so for the right-hand sides f ( t ) , we try the following:

9. f ( t ) = 2t 3 − 3t ⇒ y p ( t ) = At 3 + Bt 2 + Ct + D

10. f ( t ) = tet ⇒ y p ( t ) = ( At + B ) et 11. f ( t ) = 2sin t ⇒ y p ( t ) = Asin t + B cost

12. f ( t ) = 2e −t sin t ⇒ y ( t ) = e−t ( Acos t + B sin t )


p

 Guess Again
The homogeneous equation y′′ − 6 y′ + 9 y = 0 has characteristic equation r 2 − 6r + 9 = 0 , which has a
double root 3, 3. Hence,

yh ( t ) = c1e3t + c2 te3t .

We try particular solutions of the form:


13. f ( t ) = t cos 2t ⇒ y p ( t ) = ( At + B ) sin t + ( Ct + D ) cost

14. 3t
(
f ( t ) = te ⇒ y p ( t ) = At + Bt e
3 2
) 3t

(We can’t have any terms here dependent on terms in the homogeneous solution.)
370 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 370

15. f ( t ) = e −t + sin t ⇒ y p ( t ) = Ae −t + B sin t + C cos t

16. f ( t ) = t 4 − t 2 + 1⇒ y p ( t ) = At 4 + Bt 3 + Ct 2 + Dt + E
371 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 371

 Determining the Undetermined


17. y′ = 1 . The homogeneous solution is yh ( t ) = c , where c is any constant. By simple inspection we

observe that y p ( t ) = t is a solution of the nonhomogeneous equation. Hence, the general solution

is
y (t ) = t + c .

18. y′ + y = 1. The homogeneous solution is yh ( t ) = ce −t where c is any constant. By simple inspec-

tion we observe that y p ( t ) = 1 is a solution of the nonhomogeneous equation. Hence, the general

solution is
y ( t ) = ce −t +1.

19. y′ + y = t . yh ( t ) = ce −t , y p = At + B , y ′p = A . Substituting into the DE gives A + ( At + B ) = t .

Coefficient of t: A = 1. Coefficient of 1: A + B = 0 . Hence, A = 1, B = −1 . y p = t −1,

y = ce −t + t −1 .

20. y′′ = 1. The homogeneous solution of the equation is

yh ( t ) = c1t + c2 ,

1
where c1 , c2 are arbitrary constants. By inspection, we note that y p = t 2 is a particular
2

solution. Hence, the solution of the homogeneous equation is


1 2
y (t ) = t + c1t + c2 .
2

If you could not find a particular solution by inspection, you could try a solution of the form
y p ( t ) = At 2 + Bt + C .
21. y′′ + 4 y′ = 1. The characteristic equation is r 2 + 4r = 0 , which has roots 0, –4. Hence, the homo-
geneous solution is

yh ( t ) = c1 + c2 e −4t .

The constant on the right-hand side of the differential equation indicates we seek a particular
solution of the form y p ( t ) = A , except that the homogeneous solution has a constant solution;

thus we seek a solution of the form y p ( t ) = At . Substituting this expression into the differential

1
equation yields 4A = 1, or A = . Hence, we have a particular solution
4
372 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 372

1
y p (t ) = t ,
4
373 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 373

so the general solution is


1
y ( t ) = c + c e −4t + t .
1 2
4
22. y′′ + 4 y = 1 . The characteristic equation is r 2 + 4 = 0 , which has roots ±2i . Hence, the homoge-
neous solution is

yh ( t ) = c1 cos 2t + c2 sin 2t .

The constant on the right-hand side of the differential equation indicates we seek a particular
solution of the form y p ( t ) = A . Substituting this expression into the differential equation yields

1 1
4A = 1, or A = . We have a particular solution y p ( t ) = , so the general solution
4 4

1
y ( t ) = c1 cos 2t + c2 sin 2t + .
4
23. y′′ + 4 y′ = t . The characteristic equation is r 2 + 4r = 0 , which has roots 0, –4. Hence, the homo-
geneous solution is

yh ( t ) = c1 + c2 e −4t .

The term on the right-hand side of the differential equation indicates we seek a particular solution
of the form
y p ( t ) = At + B .

However, the homogeneous solution has a constant term so we seek a solution of the form
y p ( t ) = At 2 + Bt .

Substituting this expression into the differential equation yields


y ′ + 4 y′ = 2A + 8At + 4B = t .
1 1
Setting the coefficient of t, 1 equal to each other yields A = , B = − . Thus, the solution

8 16
1 2 1
y (t ) = c + c e + t − t .
−4t

1 2
8 16
24. y′′ + y′ − 2 y = 3 − 6t . The characteristic equation is r 2 + r − 2 = 0 , which has roots –2 and 1.
Hence, the homogeneous solution

yh ( t ) = c1e
−2t
+ c2 et .
374 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 374

The linear polynomial on the right-hand side of the equation indicates we seek a particular
solution of the form
375 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 375

y p ( t ) = At + B .

(Note that we don’t have any matches with the homogeneous solution.) Substituting this expres-
sion into the differential equation yields the equation
y′′ + y′ − 2 y = A − 2At − 2B = 3 − 6t

so A = 3 , B = 0 . Hence, we have the general solution


y ( t ) = c1e −2t + c2et + 3t .
25. y′′ + y = et + 3 . The characteristic equation is r 2 +1 = 0 , which has roots ±i . Hence, the
homogeneous solution is

yh ( t ) = c1 cos t + c2 sin t .

The terms on the right-hand side of the differential equation indicates we seek a particular
solution of the form
y p ( t ) = Aet + B .

Substituting this expression into the differential equation yields


y′′ + y = 2Aet + B = et + 3 .
1
Setting coefficients of et , 1 equal to each other, we get equations for A, B, which yield A = ,
2

B = 3 . Hence, we have the general solution


1
y ( t ) = c1 cost + c2 sin t + et + 3 .
2

26. y′′ − y′ − 2 y = 6et . The characteristic equation is r 2 − r − 2 = 0 , which has roots –1 and 2. Hence,
the homogeneous solution is
yh ( t ) = c1e2t + c2 e−t .

The exponential term on the right-hand side of the differential equation indicates we seek a
particular solution of the form
y p ( t ) = Aet .

(Note this is not linearly dependent on any of the exponential terms in the homogeneous
solution.) Substituting this expression into the differential equation we get
y′′ − y′ − 2 y = −2Aet = 6et .
373 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 373

Hence, A = −3 , and we have a particular solution

y p ( t ) = −3et ,
and hence
y ( t ) = c1e 2t + c2e −t − 3et .

27. y′′ + y′ = 6sin 2t . The characteristic equation is r 2 + r = 0 , which has roots 0 and –1. Hence, the
homogeneous solution is

yh ( t ) = c1 + c2 e−t .

The sine term on the right-hand side of the differential equation indicates we seek a particular
solution of the form
y p ( t ) = Acos 2t + B sin 2t .

Substituting into the differential equation yields


y ′′ + y ′ = ( −4 A + 2B ) cos 2t + ( −4B − 2 A ) sin 2t = 6sin 2t .

Comparing coefficients yields the equations


−4A + 2B = 0
−4B − 2A = 6,
3 6
which has the solution A , B = − . Hence, we have
=−
5 5
3 6
y p ( t ) = − cos 2t − sin 2t ,
5 5
and the general solution is
3 6
y ( t ) = c1 + c2 e−t − cos 2t − sin 2t .
5 5

y′′ + 4 y′ + 5 y = 2et . The characteristic equation of the differential equation is r + 4r + 5 = 0 ,


2
28.
which has roots −2 ± i . Hence, the homogeneous solution is

yh ( t ) = e −2t ( c1 cos t + c2 sin t ) .

The exponential on the right-hand side of the differential equation indicates we seek a particular
solution of the form
y p ( t ) = Aet .

Substituting this expression into the differential equation yields


y′′ + 4 y′ + 5 y = 10 Aet = 2et ,
374 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 374

1 1
which yields A = . Hence, we have a particular solution y p ( t ) = et , and the general solution
5 5
is given by

1
y ( t ) = e −2t ( c1 cos t + c2 sin t ) + et .
5

29. y′′ + 4 y′ + 4 y = te −t . The characteristic equation is given by r 2 + 4r + 4 = 0 , which has a double


root of –2, so the homogeneous solution is

yh ( t ) = c1e −2t + c2 te−2t .

The term on the right-hand side of the differential equation indicates we seek a particular solution
of the form
y p ( t ) = Ate + Be .
−t −t

Substituting this expression into the differential equation yields


y ′′ + 4 y ′ + 4 y = Ate −t + ( 2A + B ) e −t = te −t .
Comparing coefficients, yields equations, which we solve, getting A = 1, B = −2 . Hence, the
general solution is

y ( t ) = c1e−2t + c2 te−2t + te −t − 2e −t .

30. y′′ − y = t sin t . The characteristic equation is r 2 −1 = 0 , which has roots ±1. Hence, the
homogeneous solution is

yh ( t ) = c1et + c2 e .
−t

The term on the right-hand side of the differential equation indicates we seek a particular solution
y p ( t ) = ( At + B ) cos t + ( Ct + D ) sin t .

Differentiating this expression two times and substituting it into the differential equation yields
the algebraic equation
y ′′ − y = −2Ct sin t − 2 At cost + ( −2 A − 2D ) sin t + ( 2C − 2B ) cos t = t sin t .

Comparing terms in sin t , cos t , t sin t , t cos t , we get equations that yield
1 1
A = 0, B = − , C = − , D = 0.
2 2
Hence,
1
y ( t ) = c1et + c2 e −t − ( t sin t + cost ) .
2
375 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 375

31. y′′ + y = 12cos 2 t . The characteristic equation is r 2 +1 = 0 , which has roots ±i . Hence, the
homogeneous solution is

yh ( t ) = c1 cos t + c2 sin t .
Using the trigonometric identity
1
cos t =
2
(1+ cos 2t )

the term on the right-hand side of the differential equation yields


12cos 2 t = 6 (1+ cos 2t ) .

Hence, we seek a particular solution of the form


y p ( t ) = Acos 2t + B sin 2t + C .

Substituting this into the differential equation yields


y′′ + y = −3Acos 2t − 3B sin 2t + C = 6 + 6cos 2t .
Comparing coefficients, we get A = −2 , B = 0 , C = 6 , so the general solution is
y ( t ) = c1 cost + c2 sin t − 2cos 2t + 6 .

32. y′′ − y = 8tet . The characteristic equation is r 2 −1 = 0 , which has roots ±1. Hence, the homoge-
neous solution is

yh ( t ) = c1et + c2 e−t .

The term on the right-hand side of the differential equation indicates we seek a particular solution
y p ( t ) = Atet + Bet ,

but one term in the homogeneous solution is linearly dependent on this term, so we seek

( )
y p ( t ) = et At 2 + Bt .

Substituting this expression into the differential equation yields


y ′′ − y = 4 Atet + ( 2 A + 2B ) et = 8tet ,
which gives the two equations 4A = 8 , 2A + 2B = 0 , which gives A = 2 , B = −2 . Hence, the
general solution is

y ( t ) = c1et + c2 e−t + 2tet ( t −1) .


376 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 376

33. y′′ − 4 y′ + 4 y = te 2t . The characteristic equation of the differential equation is r 2 − 4r + 4 = 0 ,


which has a double root of 2. Hence, the homogeneous solution is

yh ( t ) = c1e2t + c2 te2t .

The term on the right-hand side of the differential equation indicates we seek a particular solution
of the form
y p ( t ) = Ate2t + Be2t ,

but both terms are linearly dependent with terms in the homogeneous solution, so we choose
() 3 2t
y p t = At e + Bt e .
2 2t

Differentiating and substituting this expression into the differential equation yields the algebraic
equation
y ′′ − 4 y ′ + 4 y = e 2t ( 0 + 0 + 6 At + 2B ) = te 2t .
1
Comparing coefficients, we get A = , B = 0 . Hence, the general solution is
6
1
y ( t ) = c1e 2t + c2te 2t + t 3e 2t .
6
34. y′′ − 4 y′ + 3y = 20cost . The characteristic equation of the differential equation is r 2 − 4r + 3 = 0 ,
which has roots 1, 3. Hence,
yh ( t ) = c1et + c2 e3t .

The term on the right-hand side of the differential equation indicates we seek
y p ( t ) = Acost + B sin t .

Substituting this expression into the differential equation yields


y ′′ − 4 y ′ + 3y = ( 4 A + 2B ) sin t + ( 2 A − 4B ) cos t = 20cos t .
Comparing coefficients yields A = 2 , B = −4 . Hence,
y ( t ) = c1et + c2 e3t + 2cost − 4sin t .

35. y′′ − 3y′ + 2 y = et sin t . The characteristic equation of the differential equation is r 2 − 3r + 2 = 0 ,
which has roots 1, 2. Hence,

yh ( t ) = c1et + c2 e2t .

The term on the right-hand side of the equation indicates we seek a particular solution
377 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 377

y p ( t ) = Aet cos t + Bet sin t .


378 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 378

Differentiating and substituting this expression into the equation yields


y ′′ − 3y ′ + 2 y = ( A − B ) et sin t + ( − A − B ) et cos t = et sin t .
1 1
Comparing coefficients, we find A = , B = − , yielding the general solution

2 2
1
y ( t ) = c1et + c2e 2t + et ( cost − sin t ) .
2
36. y′′ + 3y′ = sin t + 2cos t . The characteristic equation is r 2 + 3r = 0 , which has roots 0, –3. Hence,
the homogeneous solution is

yh ( t ) = c1 + c2 e−3t .

The sine and cosine terms on the right-hand side of the equation indicate we seek a particular
solution of the form
y p ( t ) = Acost + B sin t .

Substituting this into the equation yields


y ′′ + 3y′ = ( − A + 3B ) cos t + ( −B − 3A ) sin t = sin t + 2cost .
1 1
Comparing terms, we arrive at ( − A + 3B ) = 2 , ( −B − 3A ) = 1 , yielding A = − , B = . From
2 2
this, that the general solution is

1
y ( t ) = c + c e −3t + ( sin t − cost ) .
1 2
2
37. y′′′ − 4 y′′ = 6t

(1) Find yh: r3 − 4r2 = 0 ⇒ r2(r − 4) = 0 ⇒ r = 0, 0, 4

∴ yh = c1 + c2t + c3e4t

(2) Find yp: yP = t2(At + B) = At3 + Bt2

y p′ = 3At2 + 2Bt

y p′′ = 6At + 2B

y p′′′ = 6A

y p′′′ − 4 y p′′ = 6A − 4(6At + 2B) = −24At + 6A − 8B = 6t

coefficient of t: −24A = 6, coefficient of 1: 6A − 8B = 0


379 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 379

1 ⎛ 1⎞ 3 3
so that A = − , and 8B = 6 − = − so that B = − .
4 ⎜ ⎟
⎝ 4⎠ 2 16

1 3
Hence yp = − t 3 − t 2
4 16

1 3
(3) y(t) = y + y = c + c t + c e 4t − t 3 − t 2
h p 1 2 3
4 16
38. y (3) − 3y′′ + 3y′ − y = et

(1) Find yh: r3 − 3r2 + 3r − 1 = 0 (characteristic equation)

f(r) =− r3 −3r2 + 3r − 1 = 0

f(1) = 1 − 3 + 3 − 1 = 0 so r = 1 is a root.

By long division, we obtain

r3 − 3r2 + 3r − 1 = (r − 1)(r2 − 2r + 1) = (r − 1)3 Triple root r = 1, 1, 1

yh = c1et + c2tet + c3t2et

(2) Find yp: yp = t3(Aet) = At3et

y p′ = At 3et + 3At 2 et

y p′′ = At 3et + 3At 2 et + 3At 2 et + 6 Atet = At 3et + 6 At 2 et + 6Atet

(3) 3 t 2 t 2 t t 2 t t t t
y p = At e + 3 At e + 3 At e + 6 Ate + 3 At e + 6 Ate + 6 Ate + 6 Ae

p − 3y′′
y (3) p + 3y′p − y p = At 3et + 9At 2 et +18Atet + 6Aet

− 3At 3et −18At 2 et −18Atet + 0et


+ 3At 3et + 9At 2 et + 0tet + 0et

− At 3et + 0t 2 et + 0tet + 0et


1
6Aet = et so that A =
6
1 3 t
Thus yp = t e
6

t t 2 t 1 3 t
(3) ()
y t = yh + y p = c1e + c2 te + c3t e + t e
6
380 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 380

39. y (4) − y = 10

(1) Find yh:


r4 − 1 = 0 (r 2 +1)(r 2 −1) = (r 2 +1)(r −1)(r +1) r = ±i, ± 1

yh = c1 cos t + c2 sin t + c3et + c4e−t


(2) Find yp:
yp = A, so that y ′ = y ′ = y ′′′ = y (4) = 0
p p p p

⇒ y (4) − y = 0 − A = 10 ⇒ A = −10 ⇒ y = −10


p p p

(3) y(t) = yh + y p = c1 cost + c2 sin t + c3et + c4 e −t −10

40. y′′′ = y′′ ⇒ y′′′ − y ′ = 0

(1) Find yh: r3 − r2 = 0 ⇒ r2(r − 1) = 0 r = 0, 0, 1

(2) There is no yp because the DE is homogeneous.

(3) y(t) = c1 + c2t + c3et

 Initial-Value Problems
41. y′′ + y′ − 2 y = 3 − 6t, y(0) = −1, y′(0) = 0

(1) Find yh:

r2 + r − 2 = 0 ⇒ (r − 1)(r + 2) = 0 ⇒ r = 1, −2

∴ yh = c1et + c2e−2t
(2) Find yp: yp = At + B, y ′p = A, y ′p′ = 0 ⇒ y ′p′ + y ′p − 2 y p = A − 2( At + B) = 3 − 6t

coefficient of t: −2A = −6, coefficient of 1: A − 2B = 3 ⇒ A = 3, B = 0


∴ yp = 3t
(3) y(t) = yh + yp = c1et + c2e−2t + 3t; y′ = c1et − 2c2 e −2t + 3

y(0) = −1 ⇒ c1 + c2 = −1; y ′(0) = 0 ⇒ c1 − 2c2 + 3 = 0

5
c + c = −1 c =−
1 2 1
3

2
c − 2c = −3 c =
1 2 2
3
381 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 381

5 2
∴ y = − et + e −2t + 3t
3 3
382 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 382

42. y′′ + 4 y′ + 4 y = te −t , y(0) = −1, y′(0) = 1

(1) Find yh:

r2 + 4r + 4 = 0 ⇒ (r + 2)2 = 0 ⇒ r = −2, −2

∴ yh = c1e−2t + c2te−2t = (c1 + c2t)e−2t

(2) Find yp:


yp = e−t(At + B) ⇒ y′p = −e −t ( At + B) + Ae −t

⇒ y′′p = e −t ( At + B) − Ae −t − Ae −t = e −t ( At + B) − 2 Ae −t
So y′p′ + 4 y′p + 4 y p = e −t ( At + B) − 2 Ae−t + 4(−e −t ( At + B) + Ae −t ) + 4e −t ( At + B)
= e −t ( At + 2A + B)

This gives A = 1, 2A + B = 0 and so A = 1 and B = −2.

Therefore yp = e−t(t − 2).


(3) y = yh + yp = c1e−2t + c2te−2t + e−t(t − 2)
y ′ = −2c1e−2t + c2 e −2t − 2c2 te−2t − e −t (t − 2) + e −t

y(0) = −1 ⇒ c1 − 2 = −1 c1 = 1

y′(0) = 1⇒ −2c1 + c2 + 2 +1 = 1 c2 = 0

∴ y(t) = e−2t + e−t(t − 2)


43. y′′ + 4 y = t, y(0) = 1, y′(0) = −1

(1) Find yh: r2 + 4 = 0 ⇒ r = ± 2i ⇒ yh = c1 cos 2t + c2 sin 2t


(2) Find yp: yp = At + B, y ′p = A, y ′p′ = 0

∴ y ′p′ + 4 y p = 4( At + b) = 4At + 4B = t

1
coefficient of t: 4A = 1, coefficient of 1: 4B = 0 ⇒ A = ,B = 0
4
1
∴ yp = t
4
1 1
(3) y = yh + yp = c1 cos 2t + c2 sin 2t + t , y′ = −2c1 sin 2t + 2c2 cos 2t +
4 4

1 5 5
y(0) = 1 ⇒ c1 = 1; y′(0) = −1 ⇒ 2c2 + = −1 ⇒ 2c2 = − ⇒ c2 = −
4 4 8

5 1
∴ y(t) = cos 2t − sin 2t + t
8 4
383 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 383

44. y′′ + 2 y′ + y = 6cost, y(0) =1, y′(0) = −1

(1) Find yh: r2 + 2r + 1 = 0 ⇒ (r + 1)2 = 0 ⇒ r = −1, −1

yh = c1e−t + c2te−t
(2) Find yp: yp = A cos t + B sin t , y ′p = − Asin t + B cos t, y′p′ = − Acos t − B sin t

⇒ y′p′ + 2 y′p + y p = − Acos t − B sin t


+ 2(B cos t − Asin t)
+ Acost + B sin t

2B cos t − 2A sin t = 6 cost

coefficient of cos t: 2B = 6, coefficient of sin t: −2A = 0 ⇒ A = 0, B = 3

∴ yp = 3 sin t
(3) y = yh + yp = c1e −t + c2 te−t + 3sin t, y′ = −c1e −t − c2 te−t + c2 e−t + 3cos t

y(0) = 1 ⇒ c1 = 1; y′(0) = −1 ⇒ −c1 + c2 + 3 = −1 ⇒ c2 = −3

∴ y(t) = e −t − 3te −t + 3sin t

45. 4 y′′ + y = cos 2t, y(0) = 1, y′(0) = 0


1 1 ⎛1 ⎞ ⎛1 ⎞
(1) Find yh: 4r2 + 1 = 0 ⇒ r2 = − ⇒ r = ± i ⇒ yh = c cos ⎜ t ⎟ + c sin ⎜ t ⎟
1 2
4 2 ⎝2 ⎠ ⎝2 ⎠

(2) Find yp: yp = A cos 2t + B sin 2t, y ′p = −2 Asin 2t + 2B cos 2t, y ′p′ = −4 Acos 2t − 4B sin 2t

⇒ 4 y ′p′ + y p = −16 Acos 2t −16B sin 2t + Acos 2t + B sin 2t

= −15Acos 2t −15B sin 2t = cos 2t

coefficient of cos 2t: −15A = 1, coefficient of sin 2t: −15 B = 0


1
⇒ A= − ,B=0
15
1
∴ yp = − cos 2t
15
384 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 384

⎛1 ⎞ ⎛1 ⎞ 1
(3) y = yh + yp = c1 cos ⎜ t ⎟ + c2 sin ⎜ t ⎟ − cos 2t
⎝2 ⎠ ⎝ 2 ⎠ 15
1 ⎛ 1 ⎞ 1 ⎛ 1 ⎞ 2
y′ = − c sin t + c cos t + sin 2t
2
1 ⎜2 ⎟ 2 2 ⎜ 2 ⎟ 15
⎝ ⎠ ⎝ ⎠
1 16
y(0) = 1 ⇒ c1 − = 1 ⇒ c1 =
15 15

y ′(0) = 0 ⇒ c2 = 0

16 ⎛1 ⎞ 1
∴ y(t) = cos ⎜ t ⎟ − cos 2t
15 ⎝ 2 ⎠ 15

46. y′′ + 9 y = cos3t, y(0) =1, y′(0) = −1

(1) Find yh: r2 + 9 = 0 r = ± 3i

yh = c1 cos 3t + c2 sin 3t

(2) Find yp: yp = t(A cos 3t + B sin 3t)


y′p = t(−3Asin 3t + 3B cos3t) + ( Acos3t + B sin 3t)

y ′p′ = t(−9 Acos 3t − 9B sin 3t) + (−3Asin 3t + 3B cos3t) − 3Asin 3t + 3B cos 3t


y ′p′ + 9 y p = (9 A − 9A)t cos3t + (9B − 9B)t sin 3t + (6B) cos 3t + (−6 Asin 3t) = cos 3t

coefficient of cos 3t: 6B = 1, coefficient of sin 3t: −6A = 0


1 1
⇒ A = 0, B = so that yp = t sin 3t
6 6

1
(3) y = yh + yp = c1 cos 3t + c2 sin 3t + t sin 3t ,
6
1 1
y′ = −3c1 sin 3t + 3c2 cos 3t + t cos3t + sin 3t
2 6
1
y(0) = 1 ⇒ c1 = 1; y′(0) = −1 ⇒ 3c2 = −1 ⇒ c2 = −
3

1 1
∴ y(t) = cos3t − sin 3t + t sin 3t
3 6
383 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 383

47. y′′ − 3y′ + 2 y = 4e −t , y(0) = 1, y′(0) = 0

(1) Find yh:

r2 − 3r + 2 = 0 ⇒ (r − 1)(r − 2) = 0 ⇒ r = 1, r = 2, so that yh = c1et + c2e2t


(2) Find yp:
yp = Ae−t, y ′ = − Ae −t , y ′ = Ae−t
p p
2 2
y′p′ − 3y′p + 2 y p = Ae −t − 3(− Ae −t ) + 2 Ae −t = 6 Ae −t = 4e −t so that A = ⇒ yp = e −t
3 3

2 −t 2
(3) y = yh + yp = c1et + c2e2t + e , y′ = c1et + 2c2 e 2t − e −t
3 3
2 1
y(0) = 1 ⇒ c1 + c2 + =1 c 1 + c2 = c1 = 0
3 3
⇒ ⇒
2 2 1
y′(0) = 0 ⇒ c + 2c − =0 c + 2c = c =
1 2 1 2 2
3 3 3

1 2t 2 −t
Thus y(t) = e + e .
3 3
48. y′′ − 4 y′ + 3y = e −t + t, y(0) = 0, y′(0) = 0

(1) Find yh:

r2 − 4r + 3 = 0 ⇒ (r − 1)(r − 3) = 0 so that yh = c1et + c2e3t


(2) Find yp:
yp = Ae−t + Bt + C, y ′p = − Ae −t + B, y′′p = Ae −t

Thus y′p′ − 4 y′p + 3y p = Ae −t − 4(− Ae −t + B) + 3( Ae −t + Bt + C)

= Ae −t − 4(− Ae −t + B) + 3( Ae −t + Bt + C)

= 8Ae−t + 3Bt − 4B + 3C = e−t + t

⇒ 8A = 1, 3B = 1, −4B + 3C = 0
1 1 4 4
⇒A = ,B= , C= B= .
8 3 3 9
1 1 4
Thus, yp = e −t + t + .
8 3 9
384 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 384

(3) Find y:
1 1 4
y = yh + yp = c1et + c2e3t + e −t + t + .
8 3 9
1 1
y′ = c1et + 3c2e3t − e −t + .
8 3
1 4 41 3
y(0) = 0 ⇒ c1 + c2 + + =0 c1 + c2 = − c1 = −
8 9 72 4
⇒ ⇒
1 1 5 13
y′(0) = 0 ⇒ c1 + 3c2 − + =0 c1 + 3c2 = − c2 =
8 3 24 72
3 13 1 1 4
Therefore, y(t) = − et + e3t + e −t + t + .
4 72 8 3 9

49. y′′ − y′ − 2 y = 4 cos 2t, y(0) = 0, y′(0) = 0

(1) Find yh:

r2 − r − 2 = 0 ⇒ (r + 1)(r − 2) = 0 so r = 2, −1

yh = c1e−t + c2e2t
(2) Find yp:
yp = A cos 2t + B sin 2t, y ′p = −2A sin 2t + 2B cos 2t

y′p′ = −4A cos 2t − 4B sin 2t

y′p′ − y′p − 2 y p = −4A cos 2t − 4B sin 2t + 2A sin 2t − 2B cos 2t − 2A cos 2t − 2B sin 2t

= (−6A − 2B) cos 2t + (2A − 6B) sin 2t = 4 cos 2t

∴ coefficient of cos 2t: −6A − 2B = 4, coefficient of sin 2t: 2A − 6B = 0


1 3
so A = 3B and B = − ⇒ A= −
5 5
3 1
Thus yp = − cos 2t − sin 2t .
5 5
385 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 385

3 1
(3) y = yh + yp = c1e−t + c2e2t − cos 2t − sin 2t
5 5

6 2
y′ = −c1e−t + 2c2e2t + sin 2t − cos 2t
5 5
3 3 4
y(0) = 0 ⇒ c1 + c2 − =0 c1 + c2 = c1 =
5 5 15
⇒ ⇒
2 2 1
y′(0) = 0 ⇒ −c1 + 2c2 − = 0 −c1 + 2c2 = c2 =
5 5 3

4 −t 1 2t 3 1
∴ y(t) = e + e − cos 2t − sin 2t
15 3 5 5
50. y′′′ − 4 y′′ + 3y′ = t 2 , y(0) = 1, y′(0) = 0, y ′(0) = 0

(1) Find yh:


r3 − 4r2 + 3r = 0 ⇒ r(r2 − 4r + 3) = 0
r(r − 3)(r − 1) = 0, so r = 0, 1, 3
yh = c1 + c2et + c3e3t

(2) Find yp:

yp = t(At2 + Bt + C) = At3 + Bt2 + Ct


y′p = 3At2 + 2Bt + C, y′p′ = 6At + 2B, y′p′′ = 6A

y ′p′′ − 4 y′p′ + 3y ′p = 6A − 24At − 8B + 9At2 + 6Bt + 3C = t2

coefficient of t2: 9A = 1, coefficient of t: −24A + 6 B = 0,

coefficient of 1: 6A − 8B + 3C = 0
1 4 26
A= ,B= , C=
9 9 27
1 3 4 2 26
∴ yp = t + t + t
9 9 27
1 4 26
(3) y = yh + yp = c1 + c2et + c3e3t + t3 + t2 + t
9 9 27

Using this general solution and the initial conditions we obtain:


161 2 t 8 3t 1 3 4 2 26
y(t) = − e − e + t + t + t
81 3 81 9 9 27
386 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 386

51. y(4) − y = e2t, y(0) = y′(0) = y ′(0) = y ′(0) = 0

(a) Find yh:

r4 − 1 = 0 ⇒ (r2 + 1)(r2 − 1) ⇒ r = ± i, ± 1

yh = c1 cos t + c2 sin t + c3et + c4e−t


(2) Find yp:
yp = Ae2t, y′p = 2Ae2t, y′′p = 4Ae2t, y′′′p = 8Ae2t, y (4)
p = 16Ae2t

1 1
Thus y (4) − y = 16Ae2t − Ae2t = e2t ⇒ 15Ae2t = e2t ⇒ A = so that y = e 2t .
p p p
15 15
1 2t
(3) y = yh + yp = c1 cos t + c2 sin t + c3et + c4e−t + e
15
2
y′ = −c1 sin t + c2 cos t + c3 et − c4e−t + e 2t
15
4 2t
y′′ = −c1 cos t − c2 sin t + c3 et + c4e−t + e
15
8 2t
y′′′ = c1 sin t − c2 cos t + c3 et − c4e−t + e
15
1
y(0) = 0 ⇒ c1 + c3 + c4 + =0
15
2
y ′(0) = 0 ⇒ c + c − c + =0
2 3 4
15
4
y ′′(0) = 0 ⇒ −c + c + c + =0
1 3 4
15
8
y ′′′(0) = 0 ⇒ −c + c − c + =0
2 3 4
15
From these 4 equations in 4 unknowns, we obtain (by the methods of Chapter 3),
1 1 1 1
c1 = , c2 = , c3 = − and c4 =
10 5 4 12
1 1 1 1 1
∴ y(t) = cos t + sin t − et + e −t + e 2t
10 5 4 12 15
52. y(4) = et, y(0) = 1, y′(0) = 0, y′′(0) = 0 , y′′′(0) = 0

(1) Find yh:


387 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 387

r4 = 0 ⇒ r = 0 (multiplicity 4)

yh = c1 + c2t + c3t2 + c4t3


388 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 388

(2) Find yp:


yp = Aet, y′p = Aet, y p′ = Aet, y′′′p = Aet, y (4)
p = Aet

y (4)
p = Aet = et ⇒ A = 1 so that yp = et

(3) y = yh + yp = c1 + c2t + c3t2 + c4t3 + et


y′ = c2 + 2c3t + 3c4t2 + et

y′′ = 2c3 + 6c4t + et

y′′′ = 6c4 + et

y(0) = 1 ⇒ c1 + 1 = 1 ⇒ c1 = 0
y ′(0) = 0 ⇒ c2 +1 = 0 ⇒ c2 = −1
1
y ′′(0) = 0 ⇒ 2c +1 = 0 ⇒ c = −
3 3
2
1
y ′′′(0) = 0 ⇒ 6c +1 = 0 ⇒ c = −
4 4
6
1 1
∴ y(t) = −t − t 2 − t 3 + et
2 6

⎛t⎞
53. 4 y′′ + y = t − cos ⎜ ⎟
⎝2⎠
⎛1 ⎞ ⎛1 ⎞
Find yh: yh = c1 cos ⎜ t ⎟ + c2 sin ⎜ t ⎟
⎝2 ⎠ ⎝2 ⎠
Find yp: y p1 = At + B

⎛ ⎛t⎞ ⎛ t ⎞⎞
y p2 = t ⎜ c cos ⎜ ⎟ + D sin ⎜ ⎟ ⎟
⎝ ⎝2⎠ ⎝ 2 ⎠⎠
⎛t⎞ ⎛t⎞
∴ y p (t) = y p1 + y p2 = At + B + Ct cos ⎜ ⎟ + Dt sin ⎜ ⎟
⎝2⎠ ⎝2⎠

54. y′′′ − y′′ = t 2 + et

Find yh:

r3 − r2 = 0 ⇒ r2(r − 1) = 0 ⇒ r = 0, 0, 1

∴ yh = c1 + c2t + c3et
389 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 389

Find yp:
y p1 = t 2 ( At 2 + Bt + C), y p2 = t(Det )

4 3 2 t
∴ ()
y p t = y p1 + y p2 = At + Bt + Ct + Dte

55. y′′ − 5y′ + 6 y = cost − tet

Find yh:

r2 − 5r + 6 = 0 ⇒ (r − 2)(r − 3) = 0 ⇒ r = 2, 3

∴ yh = c1e2t + c2e3t

Find yp:
y p1 = Acost + B sin t, y p2 = et (Ct + D)

∴ y p (t) = y p1 + y p2 = Acos t + B sin t + et (Ct + D)

56. y(4) − y = tet + sin t

r4 − 1 = 0 ⇒ (r2 + 1)(r2 − 1) = 0 ⇒ r = ± i, ±1

yh = c1 cos t + c2 sin t + c3et + c4e−t


y p1 = tet ( At + B) , y p2 = t(C cost + D sin t)

∴ y p (t) = et ( At 2 + Bt) + Ct cost + Dt sin t

 Judicious Superposition

57. (a) The characteristic equation is r 2 − r − 6 = 0 has roots r = 3 , –2, so the general solution is
() 3t
yh t = c1e + c2 e
−2t
.

(b) (i) Substituting y p ( t ) = Aet yields

Aet − Aet − 6Aet = et ,


1 1
which yields A = − . Hence, y p ( t ) = − et .
6 6

(ii) Substituting y p ( t ) = Ae −t yields


390 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 390

or A Ae −t + Ae −t − 6Ae −t = e −t
1 −t
. Hence, y p ( t ) = − 4 e .
1
=−
4
389 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 389

(c) Calling L ( y ) = y′′ − y ′ − 6 y we found in part (b) that

⎛ 1 ⎞ ⎛ 1 ⎞
L ⎜ − et ⎟ = et , and L ⎜ − e −t ⎟ = e −t .
6 4
⎝ ⎠ ⎝ ⎠
1
Multiplying each equation by and using basic properties of derivatives yields
2

⎛ 1 ⎞ 1 ⎛ 1 ⎞ 1
L ⎜ − et ⎟ = et , and L ⎜ − et ⎟ = e −t
12 2 8 2
⎝ ⎠ ⎝ ⎠
and
⎛ 1
⎝ 12
1 ⎞ 1
L ⎜ − et − e −t ⎟ = et + e −t = cosh t .
8 ⎠ 2
( )
Hence, a solution of y ′ − y′ − 6 y = cosh t is

1 1
yp
(t ) = − e
t

e
−t
.
12 8

 Wholesale Superposition
58. We first solve the equation
tn
y′ + y =
n!

first getting the homogeneous


yh ( t ) = ce −t .

To find a particular solution, we try


( n)
() n n−1

yp t = An t + An−1t + … + A1t + A0 .

Substituting this into the equation yields


tn
⎡ nA t n−1 + ( n − )A (
t n−2 +…+ A + A t n + A t n−1 +…+ A t + A =

) .
n 1 n−1 1 n n−1 1 0
n!
⎣ ⎦

Comparing coefficients, we have


390 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 390

1
An =
n!
−1
An−1 =
( −1)!
n
1
A =
n−2
( n − 2 )!
−1
An−3 = ,
( − 3 )!
n

and so on. Hence, we have


n n−1 n−2
( n) t t t
yp = − + −… .
n! ( n −1)! ( n − 2 )!

Further, we have

y (p0 ) ( t ) = 1
y (p1) ( t ) = t −1
t2
y (p2 ) ( t ) = − t +1
2!
t3 t2
y (p ) ( t ) = − + t −1
3

3! 2!
4
t t3 t2
y (p ) ( t ) = − + − t + 1
4

4! 3! 2!
… … … …
n n−1 n−2
t t t
y (pn ) ( t ) = − + −…
n! ( n −1)! ( n − 2 )!

By superposition, the sum of these solutions is a solution of y′ + y = et . (We agree our discussion

is formal in the sense that we have proven superposition for finite sums.) There is a slight
problem in adding the preceding functions because the sum changes form depending on whether
we add an even or odd number of terms. We have
t2 t 4 t 2n
S 2n = y (p0 ) ( t ) + y (p1) ( t ) +… y (p2n ) ( t ) = 1+ + +…+
2! 4! ( 2n )!

t3 t 5 t 2n+1
S 2n+1 = y (p ) ( t ) + y (p ) ( t ) +… y (p ) ( t ) = t + + +…+
0 1 2n+1

3! 5! ( 2n +1)!
391 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 391

However, because the sequence S n converges, it converges to the average of the nth and
( n +1) st terms. That is,

1 1⎛ t2 t3 ⎞ 1 t
( S2n + S 2n+1 ) = ⎜1+ t + + +… ⎟ = e
2 2⎝ 2! 3! ⎠ 2

1 t
Hence, we found y p ( t ) = e .
2

 Discontinuous Forcing Functions


⎧2 0 ≤ t < 4 y(0) = y′(0) = 0
59. y′′ + y′ = ⎨
⎩1 t ≥ 4
Part 1: y1′′ + y1′ = 2 y1(0) = y1′ (0) = 0

Find (y1)h: r2 + r = 0
r(r + 1) = 0 r = 0, −1 (y1)h = c1 + c2e−t

Find (y)p: (y1)p = 2t by inspection, so

y1 = c1 + c2e−t + 2t 0 = c 1 + c2
⇒ ⇒ c2 = 2, c1 = −2
y1′ = −c2 e −t + 2 0 = −c + 2
2

∴ y1 = −2 + 2e−t + 2t
Part 2: y2′′ + y2′ = 1

y2 = c1 + c2e−t + t y2(4) = y1(4) = −2 + 2e−4 + 8 = 6 + 2e−4


y′2 = −c2e−t + 1 y2′ (4) = y1′ (4) = −2e −4 + 2

Thus, when t = 4,
6 + 2e −4 = c + c e −4 + 4
1 2
−4 −4
−2e + 2 = −c2 e +1
8 = c1 + 5 ⇒ c1 = 3
−4 −4
−2e + 1 = −c2e ⇒ −2 + e4 = −c2 ⇒c2 = 2 − e4
∴ y2 = 3 + (2e − e4)e−t + t
⎧⎪−2 + 2e −t + 2t 0≤t <4
and y(t) = ⎨
3 + (2 − e 4 )e −t + t t≥4
392 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 392

⎧cos t 0≤t ≤π y(0) = 1, y′(0) = 0


60. y′′ + 16 y = ⎨
⎩0 t >π

Part 1: y1′′ +16 y1 = cost 0≤t≤π y1(0) = 1, y1′ (0) = 0

Find (y1)h: (y1)h = c1 cos 4t + c2 sin 4t

Find (y1)p : (y1)p = A cos t + B sin t


1
⇒ B = 0, A =
( y1 )′p′ = − Acost − B sin t 15
1
Therefore, y1 = c1 cos 4t + c2 sin 4t + cos t
15

1
y1′ = −4c1 sin 4t + 4c2 cos 4t − sin t
15
1 14
y1(0) = 1 = c1 + c1 = −
15 15

y1′ (0) = 0 = 4c2 c2 = 0

14 1
∴ y1(t) = − cos 4t + cos t
15 15

Part 2: y2′′ +16 y2 = 0 t>π

y2 = c1 cos 4t + c2 sin 4t
y2′ = −4c1 sin 4t + 4c2 cos 4t
14 1 13 13
y1(π) = − + = − = y2 (π ) = −c1 c1 = −
15 15 15 15
y1′(π ) = 4 = y2′ (π ) = 4c2 c2 = 1

13
Thus y2(t) = − cos 4t + sin 4t
15

⎧ 14 1
− cos 4t + cost 0≤t ≤π

and y(t) = ⎨ 15 15

⎪− 13 cos 4t + sin 4t t >π


⎪⎩ 15
393 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 393

 Solutions of Differential Equations Using Complex Functions


61. y′′ − 2 y′ + y = 2sin t

The homogeneous solution is yh = c1et + c2tet.


For the particular solution we use y′′ − 2 y′ + y = 2eit and seek the imaginary part of the particular
solution.

We let yp = Aeit. Then y′p = iAeit and y p′ = −Aeit.

By substitution, we obtain

−Aeit − 2iAeit + Aeit = 2eit


−1
−2iA = 2 A= − =i
i

yp = ieit = i(cos t + i sin t) = i cos t − sin t

Im(yp) = cos t

∴ y(t) = yh + Im yp = c1et + c2tet + cos t


62. y′′ + 25y = 6sin t We will use y′′ + 25y = 6eit

The homogeneous solution is yh = c1 cos 5t + c2 sin 5t


For the particular solution we want Im(yp) where yp = Aeit; y ′ = iAeit and y ′ = − Aeit
p p

By substitution, we obtain

−Aeit + 25Aeit = 6eit


1
24A = 6 so that A =
4
⎛1 ⎞ 1
Im(yp) = Im ⎜ eit ⎟ = sin t
⎝4 ⎠ 4
1
∴ y(t) = c1 cos 5t + c2 sin 5t + sin t
4

We will use y′′ + 25 y = 20e


i5t
63. y′′ + 25y = 20sin 5t

The homogeneous solution is yh = c1 cos 5t + c2 sin 5t

For the particular solution we note that e5it is included in yh, so we must use an extra factor of t in
yp. We want Im(yp) where yp = Ate5it, so y′p = A(t5ie5it + e5it ) ,
394 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 394

and y′p′ = A5i(t5ie + e ) + Ai5e = A(10ie − 25te ) .


5it 5it 5it 5it 5it
395 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.4 Undetermined Coefficients 395

By substitution, we obtain
A(10ie −5i − 25te5it ) + 25Ate5it = 20e5it so that 10Ai = 20. Thus A = −2i and yp = −2ite5it.

Im(yp) = Im ( −2it(cos5t + i sin 5t) ) = −2t cos 5t

y = c1 cos 5t + c2 sin 5t− 2t cos 5t

 Complex Exponents
64. y′′ − 3y′ + 2 y = 3e 2it

The homogeneous solution is yh = c1et + c2e2t.

We seek a particular solution of the form yp = Ae2it.


Then y′ = 2iAe2it and y′′ = −4 Ae2it .

By substitution, we obtain
−4Ae 2it − 3(2iAe 2it ) + 2( Ae 2it ) = 3e 2it
−2 Ae 2it − 6Aie 2it = 3e 2it
− A(2 + 6i) = 3
−3 2 −6i −3 9
A= ⋅ = + i
2 + 6i 2 − 6i 20 20

⎛ −3 9 ⎞ ⎛ −3 9 ⎞
y p = ⎜ + i ⎟e 2it = ⎜ + i ⎟(cos 2t + i sin 2t
20 20 20 20
⎝ ⎠ ⎝ ⎠
⎛ −3 9 ⎞ ⎛ 9 3 ⎞
= ⎜ cos 2t − sin 2t ⎟ + i ⎜ cos 2t − sin 2t ⎟
20 20 20 20
⎝ ⎠ ⎝ ⎠

It can be verified directly by substitution that


3 9
Re(yp) = − cos 2t − sin 2t satisfies
20 20

y′′ − 3y′ + 2 y = 3cos 2t


9 3
and that Im(yp) = cos 2t − sin 2t satisfies
20 20

y′′ − 3y′ + 2 y = 3sin 2t

 Suggested Journal Entry

65. Student Project


395 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 395

4.5 Variation of Parameters

 Straight Stuff
1. y′′ + y = 4t

The homogeneous solutions to the equation are y1(t) = 1 and y2 (t) = e−t .
To find a particular solution of the form y p ( t ) = v1 + v2 e −t , we solve the equations

v′1 + e −t v′2 = 0
−e −t v′ = 0.
for v1′ , v′2 . This gives v1′ = 4t and v′2 = −4tet .

Integrating yields v1 ( t ) = 2t 2 and v2 ( t ) = 4et (1 − t ) .

Hence, we have a particular solution

()
y p t = y1v1 + y2 v2 = 1 2t
( ) + e ( 4e (1− t )) = 2t
2 −t t 2
− 4t + 4 .

Combining the constant term with the homogeneous solution, we write the general solution as
y(t) = c1 + c2e−t + 2t 2 − 4t .

2. y′′ − y′ = e−t

The homogeneous solutions to the equation are y1(t) = 1 and y2 (t) = et .

To find a particular solution of the form y p ( t ) = v1 y1 + v2 y2 = v1 + v2 et , we solve

v′1 + et v′2 = 0
et v′2 = e −t .
This gives v1′ = −e−t and v′2 = e−2t .

−t 1 −2t

Integrating yields v1 ( t ) = e and v2 ( t ) = − e .


2

Hence, we have a particular solution

1( )+ e ⎛− 1 e ⎞ 1 1
() −t t −2t −t −t −t

y p t = y1v1 + y2 v2 = e ⎜ ⎟=e − 2e = e .
⎝ 2 ⎠ 2
The general solution is t
y ( t ) = c1 + c2 e
396 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 396

+
1

e

.
2
397 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 397

1
3. y′′ − 2y′ + y = et , (t > 0)
t
The two linear independent solutions y1 and y2 of the homogeneous equation are

y1(t) = et and y2 (t) = tet .

Using the method of variation of parameters, we seek the particular solution


y p ( t ) = v1 ( t ) et + v2 ( t ) tet .

In order for y p (t) to satisfy the differential equation, υ1 and υ 2 must satisfy

y1v′1 + y2 v′2 = et v′te


1
t
v′2 = 0
1 .
y v′ + y v′ = et v′et (t +1)v′ = et
1 1 2 2 1 2
t
1
Solving algebraically for v1′ and v′2 we obtain v1′ ( t ) = −1 and v′2 = .
t
Integrating gives the values v1 ( t ) = −t and v2 = ln t .

Substituting these values into y p yields the particular solution

y p (t) = −tet + tet ln t .

Hence, the general solution is y(t) = c1et + c2tet + tet ln t .

4. y′′ + y = csc t

The two linearly independent solutions y1 and y2 of the homogeneous equation are
y1(t) = cost and y2 (t) = sin t .

Using the method of variation of parameters, we seek the particular solution


y p ( t ) = v1 ( t ) cost + v2 ( t ) sin t .

In order for y p (t) to satisfy the differential equation, v1 and v2 must satisfy

y1v1′ + y2 v′2 = (cost)v1′ + (sin t)v′2 = 0


y1v1′ + y2 v′2 = (−sin t)v1′ + (cost)v′2 = csct.

Solving algebraically for v1′ and v′2 we obtain v1′ ( t ) = −1 and v′2 = cot t .

Integrating gives the values v1 ( t ) = −t and v2 = ln ( sin t ) .

Substituting these values into y p yields the particular solution

y p ( t ) = −t cost + sin t ln ( sin t ) .


Hence, the general solution is y ( t ) = c1 cost + c2 sin t − t cos t + sin t ln ( sin t ) .
398 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 398

5. y′′ + y = sec t tan t

The homogeneous solutions are y1(t) = cost and y2 (t) = sin t .


We seek the solution y p ( y ) = ( cost ) v1 + ( sin t ) v2 .

(cost)v′1 (sin t)v′2 = 0


We form the system .
−(sin t)v1′ + (cos t)v′2 = sec t tan t.

Solving algebraically for v1′ and v′2 yields v1′ ( t ) = − tan 2 t and v′2 = tan t .

Integrating gives the values v1 = tan t − t and v2 = ln sect .

The particular solution is y p = (tan t − t)cost + sin t ln sec t = sin t − t cost + sin t ln sec t .

Thus, the general solution is y(t) = c1 cost + c2 sin t − t cost + sin t ln sec t .

6. y′′ − 2y′ + 2y = et sin t

The homogeneous solutions are y1(t) = et cost and y2 (t) = et sin t .


t t
To find a particular solution of the form y p ( t ) = v2 e cost + v2 e sin t ,

we solve the equations et cos tv′1 + et sin tv′2 = 0

( e cost − e sin t ) v′ + ( e sin t + e cos t ) v′ = e sin t


t t
1
t t
2
t

for v1′ and v′2 . This yields v1′ = −sin 2 t and v′2 = sin t cos t .

1 1
Integrating yields the functions v1 ( t ) =
2
( −t + cos t sin t ) and v2 ( t ) = sin 2 t .
2

Hence, a particular solution

et sin t ( sin 2 t ) =
1 1 1
y (t ) = y v + y v = et cos t ( −t + cost sin t ) + et ( sin t − t cost )
p 1 1 2 2
2 2 2

and the general solution is


1
y(t) = c1et cost + c2et sin t − tet cost .
2
399 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 399

1
7. y′′ − 3y′ + 2y =
1 + e−t
The homogeneous solutions are y1(t) = et and y2 (t) = e2t .
Hence we seek the particular solution y ( y) = e t v + e 2t v to form the system
p 1 2

et v′1 + e 2t v′2 = 0
1
et v′1 + 2e 2t v′2 = .
1+ e −t
−e −t e −2t
Solving algebraically for v1′ and v′2 yields v1′ = and v′2 = .
1+ e −t 1+ e−t
−t

The first integral is trivial; v1 = ln(1+ e ).

The second one is more difficult. However, if we perform some algebra, we can write
e −2t e −t (1+ e −t ) − e −t e −t
v′ = = = e −t − , which integrates to give v = −e −t + ln (1+ e −t ) .
2
1+ e −t 1+ e −t 1+ e −t
2

With υ1 and υ2 we have the particular solution


b g
y p = et ln 1 + e −t − et + e2t ln 1 + e−t b g
and the general solution is

y ( t ) = c1e + c2 e + e + e
t 2t
( t 2t
) ln (1+ e ) . −t

(The term −et in yp was absorbed in the homogeneous solution, giving a better form for the
solution.)
8. y′′ + 2y′ + y = e−t ln t , (t > 0)
The homogeneous solutions are y1(t ) = e−t and y2 (t ) = te−t .
We seek a particular solution y ( y) = e v1 + te v2 to form the system
−t −t
p

e−t v1′ + te −t v′2 = 0


−e −t v′1 + (e −t − te −t )v′2 = e−t ln t.

Solving algebraically for v1′ and v′2 , yields v1′ = −t ln t and v′2 = ln t .

1 1
Integrating yields v1 = − t 2 ln t + t 2 and v2 = t ln t − t .
2 4

Hence, we have a particular solution

1 2 −t 1 2 −t 2 −t 2 −t 2 −t ⎛1 3⎞
ln ln ln .
400 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 400

=− + + − = ⎜ − ⎟
yp t e t t e t e t t e t e t
2 4 ⎝2 4⎠
1
Thus the general solution is y(t) = c1e−t + c2te−t + t 2e−t (2 ln t − 3).
4
399 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 399

9. y′′ + 4 y = tan 2t
os 2t + c2 sN
yh = c1 cN in 2t
y1 y2

(− tan 2t)(sin 2t) 1 sin 2 2t 1 1− cos 2 2t 1


v1 = ∫ dt = − ∫ dt = − ∫ dt = − ∫ sec 2t − cos 2t dt
2 2 cos 2t 2 cos 2t 2

=−
1
4
( ln sec 2t + tan 2t − sin 2t )
tan 2t cos 2t 1
v2 = ∫ dt = − cos 2t
2 4

So yp = y1v1 + y2v2 = − cos 2t ( ln sec 2t + tan 2t − sin 2t ) − sin 2t cos 2t .


1 1
4 4

General solution: y(t) = c1 cos 2t + c2 sin 2t + yp.


10. y′′ + 5y′ + 6 y = cos(et )
−2t −3t
yh = c1 eN + c2 eN
y1 y2

v1 = ∫
( − cos(e ) ) e
t −3t

dt = ∫ e 2t cos(et )dt = et sin(et ) + cos(et )


−e −5t

e −2t cos(et )
v2 = ∫ dt = −∫ e3t cos(et )dt = 2sin(et ) − e 2t sin(et ) = 2et cos(et )
−e−5t

( ) (
So yp = y1v1 + y2v2 = e −2t et sin(et ) + cos(et ) + e −3t 2sin(et ) − e 2t sin(et ) − 2et cos(et ) )
= 2e −3t sin(et ) − e −2t cos(et )

General solution: y(t) = c1e−2t + c2e−3t + yp.


11. y′′ + y = sec 2 t
ost + c2 sN
yh = c1 cN in t
y1 y2

v1 = ∫ (−sec 2 t)sin dt = ∫ (−sect tan t)dt = −sect

v2 = ∫ sec2 t cost dt = ∫ sect dt = ln sect + tan t

So, yp = y1v1 + y2v2 = − cos t sect + sin t ln sect + tan t


= −1+ sin t ln sect + tan t

General solution: y(t) = cos1 cos t + c2 sin t −1+ sin t ln sect + tan t
400 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 400

et
12. y′′ − y =
t

−t
yh = c1 eNt + c2 eN
y1 y2
1 ⎛ e ⎞ −t
t
1 1 1
v = (e )dt = dt = ln t

2∫ t 2∫t
1 ⎜ ⎟
2

⎝ ⎠

1 ⎛ et ⎞ t 1 e 2t 1 t e2s

2∫ ⎝ t ⎠ 2∫ t 2 ∫ t0 s

v2 = − ⎜ ⎟ e dt = − dt = − ds

1 1 −t t e 2 s
So yp = v1y1 + v2y2 = et ln t − e ∫ ds

2 2 t0 s

General solution: y(t) = c1et + c2e−t + yp

 Variable Coefficients

13. t 2 y′′ − 2ty′ + 2y = t 3 sin t , y1(t) = t , y2 (t) = t 2

We begin by dividing the equation by t 2 , to get the proper form for using variation of parameters.
2 2
y′′ − y′ + 2 y = t sin t .
t t

Substitution verifies that y1 and y2 for a fundamental set of solution to the associated
homogeneous equation, so yh = c t + c t 2 ,

we seek a particular solution y p ( y )1= v 1t 2+ v 2t 2 ,

1 2
where v1 and v2 satisfy the conditions tv′ + t 2 v′ = 0

v′1 + 2t 2 v′2 = t sin t

Solving algebraically for v1′ and v2′ , yields v1′ = −t sin t and v2′ = sin t .

Integrating yields v1 = t cost − sin t and v2 = − cost .


Thus, y p (t) = t 2 cost − t sin t − t 2 cost = −t sin t .

Hence, the general solution of this equation is y(t) = c1t + c2t 2 − t sin t .
14. b g
t 2 y′′ + ty′ − 4y = t 2 1 + t 2 , y1(t) = t 2 , y2 (t) = t −2
401 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 401

We begin by dividing the equation by t 2 , to get the proper form for using variation of parameters:
1 4
y′′ + y′ − 2 y = 1 + t 2 .
t t
402 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 402

Substitution verifies that y1 and y2 form a fundamental set of solutions to the associated
homogeneous equation, so yh = c1t 2 + c2t −2 .

We seek a particular solution y p ( y ) = v′1t 2 + v 2t −2 ,

where υ1 and υ 2 satisfy the conditions t 2 v′ + t −2 v′ = 0


1 2

2tv′1 − 2t −3v′2 = 1+ t 2 .

1+ t 2 −t 3 (1+ t 2 )
Solving algebraically for v1′ and v2′ yields v1′ = and v2′ = .
4t 4

1 1 1 4 1 6
Integrating yields v1 = ln t + t 2 and v2 = − t − t .
4 8 16 24
2 4

1 1 t t
Thus, y (t) = t 2 ln t + t 4 − − .
p
4 8 16 24
1 1
Hence, the general solution of this equation is bg
y t = c1t 2 + c2t −2 + t ln t + t 4 .
4 12
(Notice that the term t 2 in yp can be absorbed in the homogeneous solution.)
15. (1 − t)y′′ + ty′ − y = 2(t − 1)2 e−t , y (t) = t , y t = et bg
1 2

We begin by dividing the equation by (1− t) , to get the proper form for variation of parameters
t 1
y′′ +
1− t
y′ −
1− t
y = 2 t −1 e −t b g
Susbtitution verifies that y1 and y2 form a fundamental set of solutions to the associated
homogeneous equation, so yh = c1t + c2 et

We seek a particular solution y p ( y ) = v1t + v2 e t ,

1 2
where v1 and v2 satisfy the conditions t v′ + et v′ = 0

v′1 + et v′2 = −2(t −1)e −t


Solving algebraically for v1′ and v2′ yields v1′ = 2e−t and v2′ = −2te−2t . Integrating
−t −2t ⎛ 1⎞
yields v1 = −2e and v2 = e ⎜ t + 2 ⎟.
⎝ ⎠

1 ⎛1 ⎞
Thus, y ( t ) = −2te−t + te−t + e −t = e −t −t .
p
2 ⎜2 ⎟
⎝ ⎠

Hence, the general solution of this equation is y(t) = c t + c et + e t F 1 − tI .


403 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 403


1 2 H2 K
404 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 404

16.
1
y′′ + y′ + 1−F 1 I
y = t −1 2 , y ( t ) = t −1/ 2 sin t , y ( t ) = t −1/ 2 cost
t H4t 2 K 1 2

Substitution verifies that y1 and y2 form a fundamental set of solutions to the associated
homogeneous equation, so yh = c1t −1/ 2 sin t + c2t −1/ 2 cost

We seek a particular solution y p ( y) = v 1t −1/ 2 sin t + v 2t −1/ 2 cos t ,

where v1 and v2 satisfy the conditions t −1/ 2 sin tv1′ + t −1/ 2 cos v2′ = 0

⎛ 1 −3/ 2 ⎞ ⎛ 1 ⎞
⎜− t sin t + t −1/ 2 cost ⎟ v′1 + ⎜ − t −3/ 2 cost − t −1/ 2 sin t ⎟ v′2 = t −1/ 2 .
⎝ 2 ⎠ ⎝ 2 ⎠
Multiplying through by t1/ 2 then solving for v′ and v′ : v′ = cost and v′ = −sin t.
1 2 1 2

v1 = sin t and v2 = cos t.

Thus, ( )
y p ( t ) = t −1/ 2 sin 2 t + cos 2 t = t −1/ 2 , and the general solution of this equation is

y ( t ) = yh + y p = c1t −1/ 2 sin t + c2t −1/ 2 cost + t −1/ 2 .

 Third-Order Theory
17. L( y) = y′′′ + p(t) y′′ + q(t) y′ + r(t) y = f (t)

Given yh (t) = c1 y1 + c2 y2 + c3 y3 ,

we seek y p (t) = v1 y1 + v2 y2 + v2 y3 .

Differentiating yields y ′p = v1 y1′ + v1′ y1 + v2 y2′ + v2′ y2 + v3 y3′ + v3′ y3

= = v1 y1′ + v2 y2′ + v3 y3′ (if we set y1v1′ + y2 v2′ + y3v3′ = 0 ).

Differentiating, again y ′p′ = v1′ y1′ + v1 y1′′ + v2′ y2′ + v2 y2′′ + v3′ y3′ + v3 y3′′

= v1 y1′ + v2 y2′ + v3 y3′ (if now we set y1′v1′ + y2′ v2′ + y3′ v3′ = 0 ).

Differentiating yet again: y ′p′′ = v1′ y1′′ + v1 y1′′′+ v2′ y2′′ + v2 y2′′′ + v3′ y3′′ + v3 y3′′′.

Substituting y p , y ′p , y ′p′ and y′p′′ into the L( y) = f , then regrouping all terms in v1 and v2, we see

that the coefficient of each is 0 because each yi is a solution of L(yi) = 0. Thus we are left with
y ′p′′ = y1′′v1′ + y2′′v2′ + y3′′v3′ = f .
405 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 405

This last equation, together with the two assumptions (in parentheses) that we made while
differentiating, gives a system to solve for v1′ , v2′ , v3′ :
y1v1′ + y2 v2′ + y3 v3′ = 0
y1′v1′ + y2′v2′ + y3′v3′ = 0
y1′′v1′ + y2′′v2′ + y3′′v3′ = f .
406 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 406

We use Cramer’s Rule to solve the system, then integrate to find v1 , v2 , v3 and hence, obtain a
particular solution
y p ( t ) = v1 y1 + v2 y2 + v3 y3 .

 Third-Order DEs
18. y′′′ − 2y′′ − y′ + 2 y = et
The characteristic equation ( λ −1)( λ +1)( λ − 2 ) = 0 and has roots 1, –1, and 2. The fundamental
set is y1 = et , y2 = e−t , and y3 = e2t . Hence,
yh = c1et + c2e−t + c3e2t .
By variation of parameters, we seek y p (t) = v1et + v2 e −t + v3e 2t , as in Problem 17. Hence the
system to solve is
et v′1 + e −t v′2 + e 2t v′3 = 0
et v′1 − e −t v′2 + 2e2t v′3 = 0
et v′1 + e −t v′2 + 4e 2t v′3 = et .
Using Cramer’s rule and computing the determinants yields:
⎡ 0 e −t e 2t ⎤
⎢ −t

⎢ 0 −e 2e 2t ⎥
⎡ et e −t e 2t ⎤ ⎢ ⎥
t −t 2t
⎢ ⎥ ⎢⎣ e e 4e 3e 2t 1
W = ⎢ et −e −t 2e 2t ⎥ = −6e 2t ; v1′ = = 2t
=−
W −6e 2
⎢e
t
e −t 4e 2t ⎥

⎡ et 0 e 2t ⎤
⎢ t ⎥
⎢e 0 2e 2t ⎥

⎢ ⎥
et et 4e 2t ⎥⎦ −e 4t 1
v2′ = = = e 2t
W −6e 2t 6
⎡e t
e −t 0⎤

⎢ t ⎥
⎢e −e −t 0⎥
= ⎢e
t
e −t et ⎥ −2et 1
v′ = = = e −t
3
W −6e 2t 3
1 1 1
Hence we obtain v′ = − v′ = e2t v′ = e −t .
1 2 3
2 6 3
t 1 1
Hence, v =− v = e 2t v e −t .
407 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 407

1
2
2
12 3 =−
3
1 1 1
We get a particular solution of y p (t) = − tet + et − et and the general solution is
2 12 3

1 1
y(t) = c1et + c2e−t + c3e2t − tet − et .
2 4
408 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 408

19. y′′′ + y′ = sect

Find yh: yh = r 3 + r = 0 r(r2 + 1) = 0 r = 0, ±i

yh = c1 + c2 cos t + c3 sin t
yp = v1 + v2 cos t + v3 sin t
1 cos t sin t

W = 0 −sin t cost = 1
0 − cost −sin t

0 cost sin t

0 −sin t cos t
sect − cos t −sin t cos t sin t
v1′ = = sect = sect
1 −sin t cost

v′ = ln sect + tan t

1 0 sin t
0 0 cos t

0 sec t −sin t 0 cos t ⎛ − cos t ⎞


′= = = =−

v2 1 −sin t 1⎜ ⎟ 1
1 sect sect
⎝ ⎠
v2 = −t

1 cost 0
0 −sin t 0
0 − cos t sec t −sin t 0 −sin t
′= = =− =
v3
1 1 sect sin t sect cos t
− cost
v3 = ln cos t

y(t) = c1 + c2 cost + c3 sin t + ln sect + tan t − t cos t + sin t ln cost

20. y′′ + 9 y′ = tan 3t

Find yh: r3 + 9r = 0 r(r2 + 9) = 0 r = 0 ± 3i


yh = c1 + c2 cos 3t + c3 sin 3t
yp = v1 + v2 cos 3t + v3 sin 3t
−3sin 3t
1 cos 3t sin 3t
409 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 409

3cos 3t
W = 0 −3sin 3t 3cos3t = 1 = 27
−9cos 3t −9sin 3t
0 −9cos 3t −9sin 3t
410 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 410

0 cos3t sin 3t

0 −3sin 3t 3cos 3t

tan 3t −9 cos 3t −9 sin 3t tan 3t cos3t sin 3t tan 3t tan 3t


v′ = = = (3cos 2 3t + 3sin 2 3t) =
1
27 27 −3sin 3t 3cos 3t 27 9

− ln cos3t
v1 =
27
1 0 sin 3t
0 0 3cos3t
0 tan 3t −9sin 3t −3cos 3t tan 3t 1
v2′ = = = − sin 3t
27 27 9
1
v2 = cos 3t
27
1 cos3t 0

0 −3sin 3t 0
0 −9cos 3t tan 3t −3sin 3t tan 3t 1 sin 2 3t
v3′ = = =−
27 27 9 cos3t

1 1− cos 2 3t
dt = − ( ln sec3t + tan 3t − sin 3t )
1
v3 = −
9 ∫ cos 3t 27

y = c + c cos 3t + c sin 3t −
1
ln cos3t +
1
cos 2 3t −
sin 3t
( ln sec3t + tan 3t − sin 3t )
1 2 3
27 27 27

 Method Choice
21. y′′′ − y′ = f (t)

We first find the homogeneous solution. The characteristic equation λ (λ 2 −1) = 0 has roots 0, ±
1, so the homogeneous solution is yh = c1 + c2 et + c3e−t .

(a) y′′′ − y′ = 2e −t . Because e−t is in yh, we must try yp = a te−t. The method of undetermined
coefficients is straightforward and gives a = 1, so yp = te−t and the general solution can be
−t
written y(t) = c1 + c2 et + c3e + te −t .

(b) y′′′ − y′ = sin 2 t . We cannot use undertermined coefficients on sin2 t, so we use variation
of parameters to seek a particular solution of the form yp(t) = v1 + v2et + v3e−t, with the
derivatives of v1, v2, and v3 determined from the equations
t −t
411 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 411

1v1′ + e v2′ + e v3′ = 0


t −t
0v1′ + e v2′ − e v3′ = 0
0v′1 + et v′2 + e −t v′3 = sin 2 t .

Discussion continues on next page.


412 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 412

Using Cramer’s rule (as outlined in Problem 18), we obtain


1 −t 2 1
v1′ = −sin 2 t v2′ = e sin t v3′ = et sin 2 t .
2 2

The antiderivative of υ1′ is easy to find; the other two must be left as integrals
1 1 −t 1
v1 =
2
( sin t cos t − t ) v2 = ∫ e sin 2 t dt v3 = ∫ et sin 2 t dt .
2 2

Hence, the general solution is


1 1 1
y ( t ) = c + c et + c e −t + ( sin t cos t − t ) + et e −t sin 2 t dt + e −t et sin 2 t dt .
1 2 3
2 ∫2 ∫2

(c) y ′′′ − y ′ = tan t . As in Part (b) , we must use variation of parameters to find yp, with

1v1′ + et v2′ + e −t v3′ = 0


0v1′ + et v2′ − e −t v3′ = 0
0v1′ + et v2′ + e −t v3′ = tan t .

Using Cramer’s rule (as outlined in Problem 18), to solve these equations we find
1 1
v1′ = tan t v2′ = e −t tan t v3′ = et tan t .
2 2

The antiderivative of υ1′ is easy to find; the other two must be left as integrals
1 1
v1 = ln cost v2 = ∫ e −t tan t dt v3 = ∫ et tan t dt .
2 2

Hence, the general solution is


1 1 1
y t = c + c et + c e −t + ln cos t + et e −t tan t dt + e −t et tan t dt.
() 1 2
2 ∫ 2 ∫
3
2
Parts (b) and (c) demonstrate the power of graphical methods because the algebraic
expressions for y(t) are pretty meaningless. It is easier and more informative to use DE
software to approximate solutions of this equation in ty space than it is to pursue the
analytical formula for the solution. The figures show curves for several initial conditions to
show the variety that can occur. For any IVP there would be only one solution.

Note: We used a 3D graphc DE solver with the following equations for y′′′ − y′ = f (t) :

y′ = x x′ = y
y′′ = x′ = z relisted as y′ = x
y′′′ = z′ = x + f (t) z′ = x + f (t)
413 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 413

(b) f(t) = sin2 t.

The expression for y(t) on the previous page can be further


evaluated using the identity
1
sin2 t = (1 − cos 2t),
2

but solution behavior is more easily seen on a graph of y(t).

(c) f(t) = tan t

The expression for y(t) on the previous page is even more


complicated than that for part (b); again, solution behavior is
more readily understood with a graph of y(t).

 Green’s Function Representation


22. y′′ + y = f (t)
We know that y1 = cost and y2 = sin t are the solutions of the corresponding homogeneous
equation. Their Wronksian is
cost sin t
a f
W y1, y2 (t) =
− sin t cost
= 1.

which is makes it easy to use the suggested variation of parameters formulas


− y 2 ( t )f ( t ) y1 ( t )f ( t )
v1′ = = −sin(t) f (t), v2′ = = cos(t) f (t).
W ( y1 , y2 )( t ) W ( y1 , y2 )( t )

Integrating yields
v1 = − ∫ sin ( s ) f ( s ) ds v2 = ∫ cos ( s ) f ( s ) ds.
t t
0 0

Hence, y p ( t ) = y1v1′ + y2 v2′

t t
= − cos(t) ∫ 0 sin(s) f (s)ds + sin(t) ∫ 0 cos(s) f (s)ds
t

= ∫ [ − cos(t)sin(s) + sin(t) cos(s) ] f (s)ds


0
t
= ∫ sin(t − s) f (s)ds.
0

 Green Variation
23. The homogeneous solutions are y1 = et and y2 = e−t . We seek a particular solution of the form
y p = v1 y1 + v1 y2 , where v1′ and v2′ satisfy

et v1′ + e −t v2′ = 0
e v1′ − e v2′ = f ( t ) .
414 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 414
t −t
415 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.5 Variation of Parameters 415

Adding and subtracting the equations and solving yields


1 −t 1
v1′ = e f ( t ) v′2 = − et f ( t )
2 2

1 t −s 1 t s
e f ( s ) ds e f ( s ) ds .
2 ∫0 2 ∫0
Integrating gives v1 = v2 = −

Hence, y p = v1 y1 + v2 y2
1 t t −s 1 t
= e ∫ e f ( s ) − e −t ∫ e s f ( s ) ds
2 0 2 0

t−s −t +s
⎛e −e ⎞
= ∫0 f ( s ) ⎜
t
⎟ ds
⎝ 2 ⎠
= ∫ sinh ( t − s ) f ( s ) ds.
t
0

 Green’s Follow-Up
24. From the Leibniz Rule in multivariable calculus we have the following result:
For a continuous function g(t,s),
d ⎡ ∂g ⎤
g ( t, s ) ds = lim g(r, r) +
t r
(t, s)ds ⎥ .

dt ∫ 0 ∫0
r →t
⎢⎣
t ⎦

In Problem 22, the solution of the equation y′′ + y = f (t) is


y(t) = z0
t
sin(t − s) f (s)ds

Differentiating yields
y ′ = sin ( t − t ) f ( t ) + ∫ cos ( t − s ) f ( s ) ds = ∫ cos ( t − s ) f ( s ) ds
t t

0 0

and

y ′′ = cos ( t − t ) f ( t ) − ∫ sin ( t − s ) f ( s ) ds = f ( t ) − ∫ sin ( t − s ) f ( s ) ds.


t t

0 0

Hence,

y′′ + y = f (t) − z
0
t
z t
sin(t − s) f (s)ds + 0 sin(t − s) f (s)ds = f (t) .

 Suggested Journal Entry I


25. Student Project
 Suggested Journal Entry II
26. Student Project
409 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 409

4.6 Forced Oscillations

 Mass-Spring Problems

1. x′′ + 2x′ + x = 6cost

Find xh: r2 + 2n + 1 = 0 ⇒ (r + 1)2 = 0 ⇒ r = −1

xh = c1e−t + c2te−t.
Find xp: xp = A cos t + B sin t, x′p = − Asin t + B cost, x′p′ = − Acost − B sin t

x′p′ + 2x′p + x p = − Acos t − B sin t

+ 2(B cost − Asin t)


+ Acost + B sin t
2B cost − 2 Asin t = 6cost

⇒ 2B = 6, −2A = 0 ⇒ A = 0, B = 3

xp = 3 sin t

x(t) = xh + xp = c1e−t + c2te−t + 3 sin t


⎛ π⎞
xss = 3 sin t = 3cos ⎜ t − ⎟
⎝ 2⎠
δ π radians
Amplitude C = 3; phase shift =
β 2

2. x′′ + 2x′ + 3x = cos3t

Find xh: r2 + 2r + 3 = 0 ⇒ r = −1 ± 2i

(
xh = e −t c1 cos 2t + c2 sin 2t )

Find xp: xp = A cos 3t + B sin 3t, x′p = −3Asin 3t + 3B cos3t, x′p′ = −9Acos3t − 9B sin 3t

− 9A cos 3t − 9B sin 3t)


x′p′ + 2x′p + 3x p =
+ 2(3B cos 3t − 3Asin 3t)
+ 3( Acos 3t + B sin 3t)

(−6 A + 6B) cos 3t + ( − 6 A − 6B)sin 3t = cos 3t

⇒ −6A + 6B = 1, −6A − 6B = 0
410 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 410

1 1
⇒ A= − , B = , so
12 12
411 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 411

1 1
xp = − cos3t + sin 3t
12 12
−t 1 1
x(t) = xh + xp = e ( c cos
1 )
2t + c2 sin 2t −
12
cos3t +
12
sin 3t

1 1 2 ⎛ 3π ⎞
xss = − cos3t + sin 3t = cos ⎜ 3t − ⎟
12 12 12 ⎝ 4 ⎠

; phase shift δ = π radians


2
Amplitude C =
12 β 4

3. 2x′′ + 3x = 4cos8t
3 3
Find xh: 2r2 + 3 = 0 ⇒ r2 − ⇒ r= ± i
2 2

3 3
⇒ xh = c1 cos t + c2 sin t
2 2

Find xp: xp = A cos 8t, x′p = −8Asin 8t, x′p′ = −64Acos8t

2x′p′ + 3x p = 2(−64 Acos8t)


+3( Acos8t)
−125Acos8t = 4 cos8t
⇒ −125A = 4
4
⇒A= −
125
4
⇒ xp = − cos8t
125

3 3 4
x(t) = xh + xp = c1 cos t + c2 sin t − cos8t
2 2 125
4 4
xss = − cos8t = cos(8t − π )
125 125

; phase shift δ = π radians


4
Amplitude C =
125 β 8
412 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 412

1 5
4. 2x′′ + 2x′ + x = cost
2 2

Find xh: 2r2 + 2r + 1 = 0 ⇒ r = − 1


2 2

xh = c1e −(1/ 2)t + c2te −(1/ 2)t


Find xp: xp = A cos t + B sin t, x′p = − Asin t + B cost , x′p′ = − Acost − B sin t

1
2x′′ + 2x′ + x = 2(− Acos t − B sin t)
p p p
2
+ 2(B cos t − Asin t)
1
+ ( Acos t + B sin t)
2
⎛ 3 ⎞ ⎛ 3 ⎞ 5
⎜ − 2 A + 2B ⎟ cost + ⎜ −2 A − 2 B ⎟ sin t = 2 cost
⎝ ⎠ ⎝ ⎠
3 5 3
⇒ − A + 2B = , − 2A − B=0
2 2 2
3 4
⇒ A= − , B=

5 5
3 4
xp = − cost + sin t
5 5
3 4
ce + c te − cost + sin t
−(1/ 2)t −(1/ 2)t
x(t) = xh + xp = 1 2
5 5
3 4
xss = − cost + sin t = cos(t − 2.2)
5 5

δ
Amplitude C = 1; phase shift ≈ 2.2 radians
β

5. x′′ + 2x′ + 2x = 2 cost

Find xh: r2 + 2r + 2 = 0 ⇒ r = −1 ± i
xh = e −t ( c1 cos t + c2 sin t )

Find xp: xp = A cos t + B sin t, x′p = − Asin t + B cost x′p′ = − Acost − B sin t

x′p′ + x′p + 2x p = − Acost − B sin t


+2(B cost − Asin t)
+2( Acost + B sin t)
413 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 413

( A +2B) cost + (−2 A + B)sin t = 2 cos t

⇒ A + 2B = 2, −2A + B = 0
414 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 414

2 4
⇒ A= , B=

5 5
2 4
xp = cos t + sin t
5 5
−t 2 4
x(t) = xh + xp = e (c1 cos t + c2 sin t) + cos t + sin t
5 5
2 4 2

xss = cost + sin t = cos(t −1.1)


5 5 5

2 δ
Amplitude C = ; phase shift ≈ 1.1 radians
5 β

6. x′′ + 4x′ + 5x = 2cos 2t

Find xh: r2 + 4r + 5 = 0 ⇒ r = −2 ± i
xh = e −2t (c1 cos t + c2 sin t)

Find xp: xp = A cos 2t + B sin 2t, x′p = −2 Asin 2t + 2B cos 2t, x′p′ = −4Acos 2t − 4B sin 2t

x′p′ + 4x′p + 5x p = − 4 Acos 2t − 4B sin 2t)


+ 4(2B cos 2t − 2 Asin 2t)
+ 5( Acos 2t + B sin 2t)
( A + 8B)cos 2t + (−8A + B)sin 2t = 2cos 2t

⇒ A + 8B = 2, −8A + B = 0
2 16
⇒ A= , B=
65 65
2 16
xp = cos 2t + sin 2t
65 65
2 16
x(t) = xh + xp = e −2t (c1 cost + c 2 sin t) + cos 2t + sin 2t
65 65

2 16 2
xss = cos 2t + sin 2t = cos(2t −1.4)
65 65 65

2 δ
Amplitude C = ; phase shift ≈ 0.73 radians
65 β
415 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 415

 Pushing Up
1
7. m= , b = 2.5, k = 6
4

1
x + 2.5x + 6x = 2cos 2t
4
The IVP is x + 10x + 24x = 8cos 2t, x(0) = −2, x(0) = 0.

Find xh: r2 + 10r + 24 = 0

(r + 4)(r + 6) = 0 r = −4, −6

xh = c1e−4t + c2e−6t

Find xp: xp = A cos 2t + B sin 2t


x p = −2A sin 2t + 2B cos 2t

x p = −4 Acos 2t − 4B sin 2t

x p +10x p + 24x p = −4 Acos 2t − 4B sin 2t +10(−2 Asin 2t + 2B cos 2t) + 24( Acos 2t + B sin 2t)

coeff. of cos 2t: −4A + 24A + 20B = 8 20A + 20 B = 8


coeff. of sin 2t: −4B + 24B − 20A = 0 −20A + 20B = 0
1
∴A=B=
5
1 1
xp = cos 2t + sin 2t
5 5

Therefore
1 1
x(t) = xh + xp = c1e −4t + c2e −6t + cos 2t + sin 2t
5 5

2 2
x = −4c1e −4t − 6c2e −6t − sin 2t + cos 2t
5 5
Substituting initial conditions gives
1 ⎫
x(0) = −2 = c1 + c2 +
5 ⎪ 34 23
⎬ ⇒ c1 = − , c2 =
2 5 5
x0 (0) = 0 = −4c − 6c2 + ⎪
5⎭⎪

34 −4t 23 −6t 1 1
Thus x(t) = − e + e + cos 2t + sin 2t .
5 5 5 5
416 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 416

 Pulling Down
16 1
8. m= = , b = 6, k = 16
32 2

1
x + 6x +16x = 4cos 4t
2
The IVP is x +12x + 32x = 8cos 4t, x(0) = 1, x(0) = 0 .

Find xh: r2 + 12r + 32 = 0

(r + 4)(r + 8) = 0 r = −4, −8

xh = c1e−4t + c2e−8t

Find xp: xp = A cos 4t + B sin 4t


x p = −4A sin 4t + 4B cos 4t

x p = −16 Acos 4t −16B sin 4t

x p +12x p + 32x = −16 Acos 4t −16B sin 4t + 12(−4 Asin 4t + 4B cos 4t) + 32( Acos 4t + B sin 4t)
= 8cos 4t
Coeff. of cos 4t:−16A + 32A + 48B = 8 16A + 48B = 8
Coeff. of sin 4t: −16B + 32B − 48A = 0 −48A + 16B = 0
1 3
∴A= B=
20 20
1 3
xp = cos 4t + sin 4t
20 20

Therefore
1 3
x(t) = xh + xp = c1e −4t + c2 e −8t + cos 4t + sin 4t
20 20

1 3
x = 4c1e −4t − 8c2 e −8t − sin 4t + cos 4t
5 5
Substituting intial conditions,
1 19 ⎫
1= c +c + c +c =
1 2 1 2
20 20 ⎪ 7 4
⎬ ⇒ c1 = , c2 = −
3 3⎪ 4 5
0 = −4c − 8c + c + 2c =
1 2 1 2
5 20

7 4 1 3
Thus x(t) = − e −4t − e −8t + cos 4t + sin 4t .
417 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 417

4 5 20 20
415 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 415

 Mass-Spring Again
9. (a) The mass is m = 100 kg ; gravitational force (weight) acting on the spring is

mg = 100 ( 9.8 ) = 980 newtons. Because the weight stretches the spring by

20 cm = 0.2 m , we have
980
k= = 4900 nt m .
0.20
(b) The initial-value problem for this mass is
x + 49x = 0 , x ( 0 ) = 0.40 , x ( 0 ) = 0 .

Solving we write the transient solution in polar form


x ( t ) = C cos (ω0 t − δ ) = C cos ( 7t − δ )

where the circular frequency is ω0 = 7 radians per second. Using the initial conditions
gives

x ( 0 ) = C cos δ = 0.4
x ( 0 ) = −7C sin δ = 0
or δ = 0 , C = 0.4 . Hence,
x ( t ) = 0.4cos ( 7t ) .

(c) Amplitude: C = 0.4 meter; period: T = seconds .
7

(d) If b = 500 , then


b 2 − 4mk = 250,000 − 4 (100 )( 4900 ) < 0 .

The system is underdamped.


(e) 100 x + 500x + 4900x = 0 has characteristic equation
r 2 + 5r + 49 = 0 ,
5 1
x1,2 = − ± . Hence, the general solution is
which has roots i 171
2 2
⎡ ⎛ 171 ⎞ ⎛ 171 ⎞ ⎤
x (t ) = e −5t 2
c cos t +c sin t .

⎢ 1 ⎜ ⎟ 2 ⎜ ⎟⎥
2 2
⎝ ⎠ ⎝ ⎠
Using the initial conditions x ( 0 ) = 0.4 , x ( 0 ) = 0 gives

x ( 0 ) = c1 = 0.4
171 5
416 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 416

x ( 0 ) = c2 − c1 = 0
2 2
417 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 417

2 171
which implies c1 = 0.4 , c2 = . Hence, the solution is
171

⎛ 171 2 171 171 ⎞


x ( t ) = e −5t 2 0.4cos t+ sin t .
⎜ ⎟
⎝ 2 171 2 ⎠

 Adding Forcing

10. We change the unforced equation in Problem 9 to the forced equation


100 x + 500x + 4900x = 100 cos ω f t .

(a) b 2 − 4mk = 5002 − 4 (100 )( 4900 ) < 0 , so the system is underdamped. From Equation (21)
in the text, the amplitude is a maximum when the forcing frequency is
k b2 4900 5002
ωf = − = − = 6.04 rad sec .
2 (100 )
2
m 2m2 100

(b) Given ω f = 7 , we have seen that ω0 = 7 , m = 100 , b = 500 , and hence tan δ becomes

π
infinite, so that δ = radians. Hence, by Equation (17)
2

x
ss
(t ) =
F0
(
cos ω t − δ
f
)
( 2 2
) + (bω )
2
m2 ω02 − ω f f

100 ⎛ π⎞ ⎛ π⎞
= cos ⎜ 7t − ⎟ ≈ 0.029cos ⎜ 7t − ⎟.
(100 ) ( 49 − 49 )
2 2
+ ( 500 ⋅ 7 )
2 ⎝ 2⎠ ⎝ 2⎠

See the graph for the solution of the IVP


problem. You can see that the steady
state appears to be

⎛ π⎞
xss ( t ) ≈ 0.029cos ⎜ 7t − ⎟ .
⎝ 2⎠

(c) The undamped equation is


100 x + 4900x = 100cos 7t .
The particular solution now has the form
xss ( t ) = Ct cos ( 7t − δ )

or
xss ( t ) = t ( Acos 7t + B sin 7t ) .
417 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 417

 Electric Analog

11. From Problem 10 we found the differential equation for the mechanical system to be
100 x + 500x + 4900x = 100 cos ω f t

or
x + 5x + 49x = cos ω f t .
So if R = 4 ohms , then the equivalent electrical equation (making one equation a constant multi-

ple of the other equation)


1
LQ + RQ + Q = V cos ω t
0 f
C
would be:
4 4 4 4
Q + ( 5 ) Q + ( 49 ) Q = cos ωf t .
5 5 5 5

This means we have


L = 0.80 henries
1 1
= 39.2 ⇒ C = ≈ 0.025 farads
C 39.2
V ( t ) = 0.80cos ω f t .

 Damped Forced Motion I


12. x + 8x + 36x = 72cos 6t
The given characteristic equation has roots −4 ± 2i 5 , hence in the long run the homogeneous

equation solution always decays to zero. We are only interested in a particular solution, and in
this case that solution is

x = Acos 6t + B sin 6t .
x
Differentiating and substituting into the differen-
2
tial equation gives
3 1
A= 0, B = .
2
t
Hence, the steady-state solution is given by 1 2 3

3 –1
xss ( t ) = sin 6t .
2
–2
The graph of the steady-state solution is shown.
418 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 418

 Damped Forced Motion II


13. The initial-value problem
x + 4x + 20x = 20 cos 2t , x ( 0 ) = x ( 0 ) = 0 ,

has
xh ( t ) = e −2t ( c1 cos 4t + c 2 sin 4t ) ,

and
x p ( t ) = Acos 2t + B sin 2t .

1
Substituting x p into the differential equation we find A = 1, B = , so
2

1
x p = cos 2t + sin 2t .
2

Substituting the general solution into the initial x(t)

conditions x ( 0 ) = x ( 0 ) = 0 , we find
1.5

⎛ 3 ⎞ 1
x ( t ) = −e −2t cos 4t + sin 4t + cos 2t + sin 2t . 0.5
⎜ ⎟
⎝ 4 ⎠ 2 t
1 2 3 4 5 6
Š0.5
The steady-state portion of the solution is shown.
(See figure.)
Š1.5

 Calculating Charge
14. 4Q +100Q = 10 cos 4t, Q(0) = 0, Q′(0) = 0

Find Qh : 4Q +100Q = 0 ⇒ Q + 25Q = 0 ⇒ Qh = c1 cos5t + c2 sin 5t

Find Qp : Qp has the form A cos 4t + B sin 4t.


5 5
Substitution in 4Qp +100Q p = 10 cos 4t leads to A = , B = 0 and so Qp = cos 4t.
18 18
5
Thus Q = Qh + Qp = c1 cos 5t + c2 sin 5t + cos 4t .
18
5
The initial conditions Q(0) = 0 and Q′(0) = 0 give us c1 = − , c2 = 0
18
5 5
and the solution of the IVP is Q(t) = − cos 5t + cos 4t .
18 18
419 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 419

 Charge and Current


15. Q +12Q +100Q = 12cos10t, Q(0) = 0, Q′(0) = 0

(a) Find Qh: r 2 +12r +100 = 0 ⇒ r = −6 ± 8i ⇒ Q h = e −6t (c cos8t + c sin 8t).


1 2

Find Qp: Qp has the form A cos 10t + B sin 10t.


1
Substitution in 4Qp +12Qp +100Q p = 12cos10t leads to A = 0, B =
10
1
and so Qp = sin10t.
10
1
Thus Q = Qh + Qp = e −6t (c1 cos8t + c 2 sin 8t) + sin10t.
10
1
The initial conditions Q(0) = 0 and Q(0) = 0 give us c1 = 0 c2 = −
8
1 1
and the solution of the IVP is Q(t) = − e −6t sin 8t + sin10t .
8 10
⎛3 ⎞
(b) I(t) = Q(t) = e −6t ⎜ sin 8t − cos8t ⎟ + cos10t.
⎝4 ⎠

 True/False Questions

16. True

The steady-state solution, being a particular solution, has the form A cos ωf t + B sin ωf t, where
the forcing function is Fo cos ωf t. The steady-state solution can be written in the form

xss = C cos(ωf t − δ).

Hence, the frequency of the steady-state is the same as that of the forcing function.

17. False

The amplitude of the steady-state is a function of the frequency of the forcing function.
F0
In fact, A(ω f ) = .

m2 (ω02 − ω 2f ) 2 + (bω f ) 2

We can see that A(ωf) → 0 as ωf → ∞ and that


F0
as ωf → 0.
k
420 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 420

 Beats

18. The identity


cos ( A + B ) − cos ( A − B ) = −2sin Asin B
may be used here. In this case, if A = 2t , B = t ,

and we have A + B = 3t , A − B = t . Hence,

cos3t − cost = −2sin 2t sin t .

 The Beat Goes On

19. The trigonometric identity


sin ( A + B ) − sin ( A − B ) = 2sin B cos A
may be used here. Let A = 3t and B = t . From

this we get A + B = 3t and A − B = 2 . Hence,

⎛ 3t −t ⎞ ⎛ 3t +t ⎞
sin 3t − sin t = 2sin ⎜ ⎟cos ⎜ 2 ⎟
⎝ 2 ⎠ ⎝ ⎠
= 2sin t cos 2t.

 Steady State

Note: We must be careful in finding the phase angle using the formula
B
δ = tan −1
A
B
because we don’t know in which quadrant δ lies using δ = tan −1 . The value of δ you get
A
might be π units different from the correct value. Unless you know by some other means in
which quadrant δ lies, it is best to use the two equations

C cos δ = A , C sin δ = B .

A good rule of thumb is to think of the AB plane; when both A, B are positive δ will be in the first
π
quadrant (i.e., δ between 0 and ), but when both A and B are negative δ will be in the third
2
quadrant, and so on.
421 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 421

20. x + 4x + 4x = cos t

The homogeneous solution to the equation is


() −2t −2t
xh t = c1e + c2 te .

We use the method of undetermined coefficients to find


x p ( t ) = Acos t + B sin t .
Differentiating, we get
x′p ( t ) = − Asin t + B cost
x′p′ ( t ) = − Acost − B sin t .

Substituting into the equation gives the equation


x′p′ + 4x′p + 4x p = ( − A + 4B + 4A ) cos t + ( −B − 4 A + 4B ) sin t = cos t .
3 4
Hence A = , B= with the particular solution
25 25

3 4
xp ( t ) = cos t + sin t .
25 25
Nothing in xp dies off with time; this is our
steady-state solution. Putting this in polar form
2 2
⎛ 3 ⎞ ⎛ 4 ⎞ 5 1
C = ⎜ ⎟ +⎜ ⎟ = =
⎝ 25 ⎠ ⎝ 25 ⎠ 25 5

4
δ = tan −1 ≈ 0.93 radians.
3
Hence, the steady-state solution is
x p (t) = xss ( t ) = 0.20cos ( t − 0.93) .

21. x + 2x + 2x = 2cost
The roots of the characteristic equation are −1± i . We use for the particular solution:
x p ( t ) = Acost + B sin t = C cos ( t − δ ) .
422 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 422

Another approach to x p is to note that F0 = 2 ,

ω0 = 2 , ω f = 1, m = 1, b = 2 and simply

substitute these numbers into the text solution to


find
2
xss ( t ) = cos ( t − δ )
5

where the phase angle is δ = tan −1 ( 2 ) ≈ 1.1 radi-


ans.
22. x + x + x = 4cos 3t
1 i 3
The roots of the characteristic equation are − ± . Notice that, as in previous problems, there
2 2
will be an e −t 2 involved in all of the terms in the homogeneous equation, so xh is transient, and

none of these terms will be involved in x p . We move on to find a particular solution, using the
method of undetermined coefficients. Let

x p ( t ) = Acos3t + B sin 3t .
Then we have
x′p = −3Asin 3t + 3B cos 3t , x′p′ ( t ) = −9Acos3t − 9B sin 3t .

Hence,
x′p′ + x′p + x p = ( −8A + 3B ) cos3t + ( −3A − 8B ) sin 3t = 4cos 3t .

Solving we get
32 12
A=− , B= .
73 73

In polar coordinates we have


1168 4
C = A2 + B 2 = =
73 73
⎛ 12 ⎞
δ = arctan ⎜ ⎟ ≈ 2.78 radians.
⎝ −32 ⎠

Hence, the steady-state solution is


4
x p (t) = xss ( t ) = cos ( 3t − 2.78 ) .
73
423 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 423

 Resonance
23. The differential equation is given by
x + 12x = 16cos ω t .

The circular frequency is


k
ω0 = = 12 = 2 3 radians per second,
m

the frequency is
1
f0 = 3 oscillations per second,
π

and the period of oscillations is


π
T= seconds.
3

24. If resonance exists, the input frequency ω f is the same as the natural frequency ω0 = 2 3 (see

Problem 23). Hence, we have the initial-value problem

( )
x +12x = 16cos 2 3t , x ( 0 ) = x ( 0 ) = 0 .

This equation has homogeneous solution

( )
xh ( t ) = c1 cos 2 3t + c2 sin 2 3t . ( )
To find a particular solution we seek a function of the form

( )
x p = At cos 2 3t + Bt sin 2 3t . ( )
Differentiating and substituting into the differential equation yields
16 3
A= 0, B = =4 ,
4 3 3
so the general solution is
4 3
x t = c cos 2 3t + c sin 2 3t + t sin 2 3t .
() 1 ( ) 2 ( ) 3 ( )
Substituting this into x ( 0 ) = x ( 0 ) = 0 yields c1 = 0 , c2 = 0 . Hence, the solution to the IVP is

x (t ) =
4 3
3
(
t sin 2 3t . )
424 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 424

 Ed’s Buoy
25. (a) Simple harmonic motion with
2000 125
m= = = 62.5 slugs
32 2
2π = 5 seconds ,
T=
ω0
hence

2π k
ω0 = =
5 m
or
125 4π 2
k = mω02 = × = 10π 2 .
2 25
We measure the displacement of the buoy x ( t ) from the water level with x ( t ) = 0

4 +2
corresponding to the position of the buoy with = 3 feet being above water.
2

Because the forced equation in rough seas has an amplitude of 3 feet and a period

of 7 seconds, the frequency of the forced response is . We therefore get the equation
7
2πt
62.5 x +10π 2 x = 3cos .
7

We are interested in the steady-state solution of this equation, hence we use the
method of undetermined coefficients to find a particular solution. In this case we let
2πt 2πt
xp = Acos + B sin .
7 7
49
We now differentiate and substitute into the equation yielding A = , B = 0 . Hence,
80π 2
we have

49 2π t 2π t
x p (t) = xss ( t ) = cos ≈ 0.06cos .
80π 2 7 7

[Although no friction term has been included in the preceding DE, there will in reality be
such a term, so the homogeneous solution would go to zero leaving only the oscillation
xss ( t ) .]
425 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 425

49
(b) The steady-state solution never varies more than ≈ 0.06 feet from its equilibrium
80π 2

position 3 feet above the level water line. The steady-state solution has the buoy moving
in phase with the waves so when a 3-foot wave crest hits, the buoy’s height above sea level
is approximately 3.06 feet. Thus the buoy is always at least 0.06 feet above the water
and is never submerged.

 General Solution of the Damped Forced System

26. (a) We know the form of the particular solution is


xss ( t ) = Acos ω f t + B sin ω f t .

Substituting this into the equation


mx + bx + kx = F0 cos ω f t

and simplifying, we find


( ) ( )
⎡ k − mω 2f A + bω f B ⎤ cos ω f t + ⎡ k − mω f2 B − bω f A⎤ sin ω f t = F0 cos ω f t .
⎣ ⎦ ⎣ ⎦
Setting the coefficients of the sine and cosine terms equal, yields the two equations

( k − mω ) A + bω
2
f f B = F0

( k − mω ) B − bω
2
f f
A = 0.

Solving, we obtain

A=
(
F0 k − mω f2 )
( k − mω ) 2 2
f + b ωf2 2

F0bωf
B= .
( k − mω ) 2 2
f + b 2ωf2

(b) From part (a) we have

x (t ) =
F0
(
⎡ k − mω 2 cos ω t + bω sin ω t ⎤ .)
( k − mω ) 2 ⎣
ss f f f f ⎦
2 2
f + b 2ωf

Rewriting this in polar form, yields

x
ss
(t ) =
F0
( )
cos ω t − δ ,
f
( k − mω )
2
2
f +b ω 2 2
f

bω f
with tan δ = . From this equation it can be seen that the long-term response
(
m ω02 − ω f2 )
of the system is oscillatory with the same frequency ω f as the forcing term, but with a

phase lag.
426 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 426

 Phase Portrait Recognition


27. x + 0.3x + x = cost

(C) We have damping but we also have a sinusoidal forcing term. Hence, the homogeneous
solution goes to zero and particular solutions consist of sines and cosines, which give rise to
circles in the phase plane. Therefore, starting from the origin x ( 0 ) = x ( 0 ) = 0 we get a curve that
approaches a circle from the inside.

28. x+x=0

(A) The equation models the undamped harmonic oscillator, which has circular trajectories.

29. x + x = cos t
(D) This equation has resonance so the trajectories in phase space spiral to infinity.

30. x + 0.3x + x = 0

(B) The system is unforced but damped, and hence trajectories must approach x ( 0 ) = x ( 0 ) = 0 .

 Matching 3D Graphs

31. (a) E (b) B (c) C, D (d) A, C


(e) B, D, E (f) C, D (g) A

 Mass-Spring Analysis I
32. (a) xh = 4cos 4t − 3sin 4t (b) The amplitude of xh = 5 .

(c) The amplitude (time-varying) of x p is 5t .

(d) x p = 5t sin 4t ; x p will be unchanged.

k
(e) Because ω f = ω0 , ω0 = 4 = = k , k = 16 .
m

(f) The system is in a state of pure resonance because ω0 = ω f . The mass will oscillate with
increasing amplitude.
 Electrical Version
33. (a) Qh = 4cos 4t − 5sin 4t

(b) The amplitude of the transient solution is A = 42 + 52 = 41

(c) Qs = Q p = 6t cos 4t (d) Q p = 6t cos 4t

1 1 1
(e) 4= = , C=
LC C 16
427 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 427

(f) The charge on the capacitor will oscillate with ever-increasing amplitude due to pure
resonance.
428 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 428

 Mass-Spring Analysis II
34. (a) xh = 3e −2t cos t − 2e −2t sin t

−b
(b) From the exponential function e −2t we see that = −2. Hence if m = 1, b = 4 .
2m
(c) Underdamped

(d) The amplitude (time-varying) of xh = 13e −2t .

(e) xss = x p = 2 cos ( 5t − δ )


4k −16
(f) ω f = 5 rad sec , β = 1 = so k = 5 Nt m , and ω0 = 5 rad sec . From For-
2

F0 F0
mula (19), we obtain 2= = . Therefore F = 40 Nt .

(5 − 5 )
2
+ ( 4 ⋅ 5) 800
2 2
0

 Perfect Aim

35. (a) The dart is fired straight at the target with initial velocity vo.
Let yD denote the vertical position of the dart at time t.
y D′′ = −g y D′ = −gt + c
d = x02 + y02
y D′ (0) = vo sin θ so y D′ = −gt + v0 sin θ
1
Integrating: yD = − gt 2 + (v0 sin θ )t + c
2

1
y D (0) = 0 so yD = − gt 2 + (v0 sin θ )t
2

Now consider the target. Let yT denote the vertical position of the target at time T. The
1
initial conditions are yT(0) = y0 and yT′ (0) = 0. By similar calculations, yT = y0 − gt2.
2

(b) To find the time t1 when the heights of dart and target are equal, set yT(t) = yD(t).
1 1 y 0 ⎛ y ⎞ y
Then yo − gt = v sin t −
2 2
gt so that T1 = 0
and x1 = (v cos θ ) ⎜ 0
⎟=
0

2
0
2 v0 sin t ⎝ v0 sin θ ⎠ tan θ

y0 y0
However, tan θ = = so that x 1 = x0 (i.e., the dart hits the target).
x0 x1
429 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.6 Forced Oscillations 429

(c) Substituting t1 into either equation for the height of the dart or the target at impact yields yT.
gy 02
yT = y0 −
2(v0 sin θ ) 2
y
Simplifying by using the diagram so that sin θ = 0 , we obtain
d
2
1 ⎛d ⎞
yT = y0 − g ⎜ ⎟ .
2 ⎝ v0 ⎠
428 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 428

 Extrema of the Amplitude Response


36. We write

A ωf = ( ) F0
=
F0 / m
.

( k − mω )
2 2
2
+ b 2ω 2 ⎡⎛ k ⎞ ⎤
2
+ ⎛⎜ ⎞⎟ ω2f
f f
b
2
⎢⎜ m ⎟ − ω f ⎥ ⎝ ⎠ m
⎣⎝ ⎠ ⎦

Differentiating A with respect to ω f , we find


⎛ 2ωf ⎞ ⎡ 2
⎜ m ⎟ f
ω − k − b ⎤
m 2m2
2
( )
( )
A′ ω f = ⎝ ⎠
32
× F0

(
⎡ k
) + ( mb ) ω ⎤⎥
2 2
ω2 2

⎢ − f f
⎣ m ⎦
( )
from which it follows that A′ ω f = 0 if and only if ω f = 0 or

k b2
ωf = − .
m 2m2

( )
When b 2 > 2mk , ω f is not real. Hence A′ ω f = 0 only when ω f = 0 . In this case

( )
A ωf damps to zero as ω f goes from 0 to ∞. It is clear then that the maximum of A ω f ( )
occurs when ω f = 0 and has the value
1
A(0) = .
k
When b 2 < 2mk , then ω f is real and positive. It is easy using the sign of the derivative to see

that the maximum of A ω f ( ) occurs at


k b2
ωf = − .
m 2m2

Evaluating the amplitude response at this value of ω f yields the expression

F0
Amax = .
2
b k − b
m 4m2

 Suggested Journal Entry

37. Student Project


429 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 429

4.7 Conservation and Conversion

 Total Energy of a Mass-Spring


1. x + x = 0 , x ( 0 ) = 1 , x ( 0 ) = −4

1 2 1 2
The total energy of the system is E = mx + kx . Here m = 1, k = 1 , so
2 2
1 2 1 2
E= x + x .
2 2
Because the system is conservative it does not change over time. Initially we have x ( 0 ) = 1 ,
x ( 0 ) = −4 , so the initial energy of this system is

1 1 17
E=
2
( −4 ) + (1) = ,
2

2 2

which remains constant in time.

 Nonconservative Mass-Spring System


2. x + 2x + 26x = 0 , x ( 0 ) = 1 , x ( 0 ) = 4

(a) The solution of the IVP is


x ( t ) = e −t ⎡⎣sin ( 5t ) + cos ( 5t ) ⎤⎦ .

π
At time t = we have
5
⎛π⎞
x ⎜ ⎟ = −e −π 5 .
⎝5⎠

Also
x ( t ) = −e −t ⎡sin
⎣ ( 5t ) + cos ( 5t ) + e ⎡⎣5cos ( 5t ) − 5sin ( 5t ) ,
−t

so
⎛π⎞
x ⎜ ⎟ = e −π 5 − 5e −π 5 = −4e−π 5 .
⎝5⎠

(b) Because m = 1, k = 26 , we have


2 2
⎛π⎞ ⎡ ⎛ π ⎞⎤ 1 ⎡ ⎛ π ⎞⎤
E
1
= m x + k x
1
= 16e −2π ( 5
) + 1 ( 26e −2π 5
) = 21e −2π 5
.
⎜ 5 ⎟ 2 ⎢ ⎜ 5 ⎟⎥ 2 ⎢ ⎜ 5 ⎟⎥ 2 2
⎝ ⎠ ⎣ ⎝ ⎠⎦ ⎣ ⎝ ⎠⎦
430 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 430

(c) Because the initial energy of the system was

E (0) =
1
2
1 1 1
( ) (
m ⎣⎡ x ( 0 )⎦⎤ + k ⎣⎡ x ( 0 )⎦⎤ = 1⋅ 42 + 26 ⋅12 = 21 ( joules or ergs )
2

2
2

2 2
)

the energy loss = 21− 21e −2π 5 = 21 1− e −2π 5 . ( )


 General Formula for Total Energy in an LC-Circuit
1
3. LQ + Q = 0 , Q ( 0 ) = Q0 , I ( 0 ) = I 0
C

The total energy of this LC system is the constant value


1 1 1 1
()
E t = LQ +
2 2
Q =
2
LI 0 +
2
Q0 .
2 2C 2 2C

 Energy in an LC-Circuit
1
4. LQ + Q = 0 , Q ( 0 ) = Q0 , I ( 0 ) = I 0
C

The total energy of this LC system is the constant value

E (t ) =
1
2
LQ 2 +
1 2 1 2 1 2 1
2C
Q = LI0 +
2 2C
Q0 = 4 12 +
2
16 2
2
4 = 130 . ( ) ( )

 Energy Loss in LRC-Circuit


5. LQ + RQ + CQ = 0 , Q ( 0 ) = Q0 , I ( 0 ) = I 0
We are given L = 1 henry, R = 1 ohm, C = 4 farads, Q0 = 0 coulomb, I 0 = 2 amps. Hence, the
IVP is

Q + Q + 0.25Q = 0 , Q ( 0 ) = 0 , I ( 0 ) = 2 ,
whose solution is given by
Q ( t ) = 2te −t 2
I ( t ) = Q ( t ) = 2e −t 2 − te −t 2 = −e −t 2 ( t − 2 ) .

Hence, the initial energy is

E (0) =
1 2 1 2 1
LI +
1
Q = (1) 22 + ( 0 ) = 2 joules.
2 0 2C 0 2 8
( )
At time t the energy is

E (t ) =
1 2
2
LI ( t ) +
1 2
2C
1 1 1
( )
Q ( t ) = e −t ( t − 2 ) + 4t 2 e −t = e −t ⎡( t − 2 ) + t 2 ⎤ = e −t t 2 − 2t + 2 .
2
2

8 2 ⎣
2
⎦ ( )

(
Hence, after time t the energy loss is 2 − e −t t 2 − 2t + 2 joules. )
431 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 431

 Questions of Energy

6. x − x + x3 = 0

(a) KE =
1
( ) 1 1
x 2 , V = ∫ −x + x 3 dx = − x 2 + x 4
2 2 4
1 2 1 2 1 4
E ( x, x ) = KE + V = x − x + x
2 2 4

(b) To find the equilibrium points, we seek the solutions of


∂E ∂E
= −x + x3 = 0 , = x =0.
∂x ∂x

Solving these equations, we find three equilibrium points at ( −1, 0 ) , (0, 0) and (1, 0).
Because x = 0 for all these points, we determine which points are stable (local maxima)
by simply drawing the graph of V ( x ) (shown in part (c)).

(c) Graph of the potential energy V ( x ) is V (x)

shown. Note that V ( x ) has local minima 3


at x = ±1 and a local maxima at x = 0 .

Hence, 2

( −1, 0 ) and (1, 0 )


1

are stable points, and


( 0, 0 ) –2 –1
–0.5
1 2
x

is an unstable point.

Potential energy of x − x + x3 = 0

7. x − x − x3 = 0

(a) KE =
1
( ) 1 1
x 2 , V = ∫ −x − x3 dx = − x 2 − x 4
2 2 4
1 2 1 2 1 4
E ( x, x ) = KE + V = x − x − x
2 2 4

(b) To find the equilibrium points, we seek the solutions of the two equations
∂E ∂E
= −x − x3 = 0 , = x =0.
∂x ∂x

Solving these equations, we find one equilibrium point at ( 0, 0 ) . Because x = 0 we


432 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 432

determine if it is a stable point (local minima) or unstable point (local maxima) by simply
drawing the graph of V ( x ) shown in part (c).
433 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 433

The graph of potential energy V ( x ) is V (x)

shown. Note that V ( x ) has local maxima 1


at x = 0 , and hence ( 0, 0 ) is an unstable

equilibrium point. x
–2 –1 1 2

(c) See the figure to the right. –1

–2
Potential energy of x − x − x3 = 0

8. x − x + x2 = 0

(a) KE =
1
( 1
) 1
x 2 , V = ∫ −x + x 2 dx = − x 2 + x 3
2 2 3
1 2 1 2 1 3
E ( x, x ) = KE + V = x − x + x
2 2 3

(b) To find the equilibrium points, we seek the solutions of the two equations
∂E ∂E
= −x + x 2 = 0 , = x =0.
∂x ∂x

Solving these equations, we find two equilibrium points at ( 0, 0 ) , (1, 0 ) . Because x = 0

for both these points, we determine which points are stable (local minima) and which are
unstable (local maxima) by simply drawing the graph of V ( x ) shown in part (c).

The graph of potential energy V ( x ) is V (x)

shown. Note that V ( x ) has local minima 1


at x = +1 and a local maxima at x = 0 .
Hence, ( 0, 0 ) is an unstable point and x
–2 –1 1 2

(1, 0 ) is a stable point.


–1
(c) See figure.
–2

Potential energy of x − x + x 2 = 0

9. x + x2 = 0

1 1
(a) KE = x 2 , V = ∫ x 2 dx = x3
2 3
434 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 434

1 2 1 3
E ( x, x ) = KE + V = x + x
2 3
(b) To find the equilibrium points, we seek the solutions of the equations
∂E ∂E
= x2 = 0 , = x =0.
∂x ∂x

Solving these equations, we find one equilibrium point at


( 0, 0 ) .
To determine if the point is stable, we V (x)

note that the potential energy 3


1 3
V ( x) = x 2
3

does not have a local maxima or minima –2 –1 1 2 x

at ( 0, 0 ) , so ( 0, 0 ) is an unstable (or –1

semistable) equilibrium point.


–2
(c) The graph of V ( x ) is a simple cubic.
Potential energy of x + x 2 = 0

10. x − e x −1 = 0

(a) KE =
1 2
2
( )
x , V = ∫ −e x −1 dx = −e x − x

1
E ( x, x ) = KE + V = x 2 − e x − x
2

(b) To find the equilibrium points, we seek the solutions of the two equations
∂E ∂E
= −e x −1 = 0 , = x =0.
∂x ∂x
It is clear that the first of these equations V (x)

has no root, so the equation has no equi- 2


librium points.
1
(c) Note that the graph of the potential energy
V ( x ) does not have any local maxima or –2 –1 1 2 x

minima points, which corresponds to the –1


lack of equilibrium points found in part
–2
(b).

Potential energy of x − e x −1 = 0
435 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 435

x + ( x −1) = 0
2
11.

1 1
x 2 , V = ∫ ( x −1) dx =
2
(a) KE = x3 − x 2 + x

2 3
1 2 1 3
E ( x, x ) = KE + V = x + x − x2 + x
2 3

(b) To find the equilibrium points, we seek the solutions of the two equations
∂E ∂E
= x 2 − 2x +1 = 0 , = x =0.
∂x ∂x

Solving these equations, yields only one real equilibrium point (1, 0 ) . Because x = 0 ,

we determine if it is stable (local minima) or unstable (local maxima) by simply drawing


the graph of V ( x ) show in part (c). Thus, we find that (1, 0 ) is an unstable (or

semistable) equilibrium point.


(c) The graph of the potential energy V ( x ) V (x)

is shown. Note that V ( x ) has neither a 2

maximum nor a minimum at x = 1 , and


1
hence (1, 0 ) is an unstable (or semista-

ble) equilibrium point. x


–2 –1 1 2

–1

–2

Potential energy of x + ( x −1) = 0


2

1
12. x=
x2
1 1 1
(a) KE = x2 , V = −∫ dx =
2 x2 x
1 2 1
E ( x, x ) = KE + V = x +
2 x

(b) To find the equilibrium points, we seek the solutions of the two equations
∂E −1 ∂E
= 2 =0, = x =0.
∂x x ∂x
436 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 436

Because the first equation does not have a solution, there is no equilibrium point.
437 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 437

(c) The graph of the potential energy V ( x ) V (x)

is shown. Note that V ( x ) does not have 4

any local maxima or minima, which cor-


2
responds to the absence of equilibrium
points noted in part (b). –2 –1 1 2 x

–2

–4
Potential energy of x = 1 x 2

13. x = ( x −1)( x − 2 )

1 1 3
(a) KE = x 2 , V = − ∫ ( x −1)( x − 2 ) dx = x3 + x 2 − 2x

2 3 2
1 2 1 3 3 2
E ( x, x ) = KE + V = x − x + x − 2x
2 3 2

(b) To find the equilibrium points, we seek the solutions of the two equations
∂E ∂E
= −x 2 + 3x − 2 = 0 , = x =0.
∂x ∂x

Solving these equations, we find two equilibrium points at (1, 0 ) and ( 2, 0 ) . Because

x = 0 , we determine which points are stable (local minima) and which are unstable (local
maxima) by simply drawing the graph of V ( x ) show in part (c).

(c) The graph of the potential energy V ( x ) V (x)

is shown. Note that V ( x ) has local 2

minima at x = 1 and a local maxima at


1
x = 2 . Hence, (1, 0 ) is a stable point and

( 2, 0 ) is an unstable point.
–4 –2 2 4 x

–1

–2

Potential energy of x = ( x −1)( x − 2 )


436 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 436

 Conservative or Nonconservative?
14. x + x2 = 0 x
4
Conservative because it is of the form
2
mx + F ( x ) = 0 .

–4 –2 2 4 x
The total energy of this conservative system is
–2
1 1 1
E ( x, x ) = mx + ∫ F ( x ) dx = x 2 + x3 .
2 2 3 –4

We draw contour curves for this surface over the xx -plane to view the trajectories of the
differential equation in the xx plane.
15. x + kx = 0 x

Conservative because it has the form 4

mx + F ( x ) = 0 . 2

The total energy of this conservative system is –4 –2 2 4 x


–2
1 1 1
E ( x, x ) = mx + ∫ F ( x ) dx = x 2 + kx 2 .
2 2 2 –4

We draw contour curves for this surface over the xx -plane to view the trajectories of the differen-
tial equation in the xx plane. The trajectories of x + kx = 0 are ellipses each with height k
times its width.

16. x + x + x2 = 1 x
5
Not conservative due to the x term. The spiral
trajectories in its phase plane cannot be level

curves of any surface. x


–5 5

–5
437 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 437

17. θ + sin θ = 0 θ
4
Conservative because it is of the form
mθ + F (θ ) = 0 . 2

The total energy of this nonconservative system −π π θ


is –2

( )1 1
E θ , θ = mθ + ∫ F (θ ) dθ = θ 2 − cos θ .
2 2 –4

We can draw contour curves for this surface over the θθ -plane to view the trajectories of the
differential equation in the θθ plane.

18. θ + sin θ = 1 θ
2
Conservative because it can be written in the
form 1

mθ + F (θ ) = 0 , θ
–8 –4 4 8
where F (θ ) = sin θ −1 . The total energy is
–1

( )
1
E θ , θ = θ 2 − cosθ − θ .
2 –2

We can draw contour curves for this surface over the θθ -plane to view the trajectories of the
differential equation in the θθ plane.

19. θ + θ + sin θ = 1 θ
4
Not conservative due to the θ term. The
following phase plane portrait shows equilibria 2
along the axis. Trajectory cannot be level curves
θ
for any surface. −2π −π π 2π

–2

–4

Trajectories of a nonconservative system


438 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 438

 Time-Reversible Systems
20. (a) mx = F ( x ) . If we introduce backwards time τ = −t , then taking the derivatives, yields

dx dx dτ dx
= =−
dt dτ dt dτ

d 2x d d ⎛ dx ⎞ dτ d2x d 2x
= ( x) = − = − 2 ( −1) = 2
dt 2 dt dτ ⎜⎝ dτ ⎟⎠ dt dτ dτ

The conservative system x + F ( x ) = 0 is transformed into exactly the same equation

d2x
+ F ( x) = 0
dτ 2

in backwards time τ .
(b) The solution of the IVP
x + x = 0 , x (0) = 1 , x (0) = 0

is x ( t ) = cost . If we replace t by –t, it yields the solution x ( −t ) = cos ( −t ) = cost . Hence,

running the system backwards looks exactly like running the system forward.
(c) The solution of the IVP x = −mg , x ( 0 ) = 0 , x ( 0 ) = 100 is

1
x (t ) mgt 2 + 100t .
=−
2
1
If we replace t by –t, we get x ( −t ) = − 2 mgt −100t . Hence, the solution is not the same,
2

and the system is not time reversible.

(d) If we think of a time-reversible system as a system where equations of motion are the
same when we replace t by –t, we might make the following conclusions.

(i) yes (ii) no (iii) no

(iv) no (v) yes

 Computer Lab: Undamped Spring

21. IDE Lab

 Computer Lab: Damped Spring

22. IDE Lab


439 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 439

 Conversion of Equations
23. x + ω02 x = f ( t )

Letting x1 = x , x2 = x , we have
x1 = x2
x2 = −ω02 x1 + f ( t ) .

In matrix form, this becomes


⎡ x1 ⎤ ⎡ 0 1⎤ ⎡ x1 ⎤ ⎡ 0⎤
= + .
⎢x ⎥ ⎢ 2 ⎥⎢ ⎥ ⎢f t ⎥
⎣ 2⎦
−ω 0 x2 ()
⎣ 0 ⎦⎣ ⎦ ⎣ ⎦
g
24. θ+ sin θ = 0
L

Letting x1 = θ , x2 = θ , we have

x1 = x2
g
x2 = − sin x1 .
L
This system is not linear, so there is no matrix form.
25. ay′′ + by′ + cy = 0

Letting x1 = y , x2 = y′ , we have
x1 = x2
b
x2 = − c cx − x2 .
1 a
a

In matrix form, this becomes

⎡ x1 ⎤ ⎢⎡
0 1⎤ x
⎡ 1⎤
=
⎢x ⎥ ⎢ c b ⎥⎥ ⎢ ⎥ .
− − x
⎣ 2⎦ ⎣ 2⎦
a a
1
26. LQ + RQ + Q=0
C
dQ
Letting x1 = Q , x2 = , we have
dt
x1 = x2
1 R
x =− x − x .
440 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 440
2 1 2
LC L
440 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 440

In matrix form, this becomes

⎡ x1 ⎤ ⎢⎡
0 1⎤
⎡ x1 ⎤
⎢x ⎥ = ⎢ 1 R ⎥⎥ ⎢ ⎥ .
− − x
⎣ 2⎦ ⎣ 2⎦
LC L

27. ( )
t 2 x + tx + t 2 − n 2 x = 0

Letting x1 = x , x2 = x , we have
x1 = x2

x2 =−
(t 2
− n2 ) x − 1x .
2 1 2
t t

In matrix form, this becomes


⎡ 0 1⎤
⎡ x1 ⎤ ⎢ 2 ⎥ ⎡ x1 ⎤
⎢ ⎥ = t −n 2
1 ⎢ ⎥.

⎣ x2 ⎦ ⎢ − − ⎣ x2 ⎦
⎣ t2 t⎦

28. x + (1+ sin ω t ) x = 0

Letting x1 = x , x2 = x , we have
x1 = x2
x2 = − (1+ sin ω t ) x1 .

In matrix form, this becomes


⎡ x1 ⎤ ⎡ 0 1⎤ ⎡ x1 ⎤
= .
⎢x ⎥ ⎢ − 1+ sin ω t 0 ⎥ ⎢ ⎥
⎣ 2⎦ ⎣ ( ) ⎦ ⎣ x2 ⎦
29. (1− t ) y′′ − 2ty′ + n ( n +1) y = 0
2

Letting x1 = y , x2 = y′ , we have
x1 = x2
n ( n +1) 2t
x2 = − 2
x1 + x2 .
1− t 1− t2

In matrix form, this becomes


⎡ 0 1 ⎤
⎡ x1 ⎤ ⎢ ⎡ x1 ⎤
⎢ ⎥ = ⎢ n +1 2t ⎥⎥ ⎢ ⎥ .
x −n x
⎣ 2⎦ ⎣ 2⎦
1− t 2 1− t 2
441 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 441

In matrix form, this becomes


d4y d3y d 2 y dy
30. 4 +3 3 +2 + + 4y =1
dt dt dt 2 dt

If we introduce
x1 = y
dy
x2 =
dt
d2y
x3 = 2
dt
d3y
x4 = 3
dt

we have the differential equations


x1 = x2
x2 = x3
x3 = x4
x4 = −4x1 − x2 − 2x3 − 3x4 +1

or in matrix form
⎡ x1 ⎤ ⎡ 0 1 0 0 ⎤ ⎡ x1 ⎤ ⎡0⎤
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥
⎢ x2 ⎥ = ⎢ 0 0 1
0 ⎥ ⎢ x2 ⎥ ⎢ 0 ⎥
⎢ x3 ⎥ ⎢ 0 0 0 + .
1⎥ ⎢ x3 ⎥ ⎢ 0 ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥
⎣ x4 ⎦ ⎣ −4 −1 −2 −3⎦ ⎣ x4 ⎦ ⎣1 ⎦

 Conversion of IVPs
31. y′′ − y′ + 2 y = sin t , y ( 0 ) = 1, y ′ ( 0 ) = 1

Letting x1 = y , x2 = y′ yields
x1′ = x2 x1 ( 0 ) = 1
.
x2′ = −2x1 + x2 + sin t x2 ( 0 ) = 1
In matrix form this becomes
⎡ x1′ ⎤ ⎡ 0 1⎤ ⎡ x1 ⎤ 0⎤ ⎡ x1 ( 0 ) ⎤ ⎡1⎤

= + ; = .
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎣ x2′ ⎦ ⎣ −2 1⎦ ⎣ x2 ⎦ ⎣sin t ⎦ ⎣ x2 ( 0 ) ⎦ ⎣1⎦

32. y′′′ + ty′ + y = 1, y ( 0 ) = 0 , y ′ ( 0 ) = 1 , y ′′ ( 0 ) = 2

Letting x1 = y , x2 = y′ , x3 = y ′ yields x1′ = x2


x2′ = x3
442 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 442

In matrixx3′form,
= −x1this becomes
− tx2
x1 ( 0 )
+1
=1
x2 ( 0 )
=1 .
x3 ( 0 )
=2
443 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 443

In matrix form, this becomes


⎡ x1′ ⎤ ⎡ 0 1 0 ⎤ ⎡ x1 ⎤ ⎡ 0 ⎤ ⎡ x1 ( 0 ) ⎤ ⎡0 ⎤
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ ⎥
⎢ x2′ ⎥ = ⎢ 0 0 1⎥ ⎢ x2 ⎥ + ⎢ 0 ⎥ , ⎢⎣ x2 ( 0 ) ⎥⎦ = ⎢ 1 ⎥ .

x3′ ⎢⎣ −1 −t 0 x3 ⎢⎣1 ⎢ x3 ( 0 ) ⎥ ⎢⎣ 2

33. y′′ + 3y′ + 2z = e −t , y ( 0 ) = 0 , y ′ ( 0 ) = 1

z′′ + y + 2z = 1, z ( 0 ) = 1, z ′ ( 0 ) = 0

Letting x1 = y , x2 = y′ , x3 = z , x4 = z′ yields
x1′ = x2 x1 ( 0 ) = 0

x2′ = −3x2 − 2x3 + e−t x2 ( 0 ) = 1


.
x3′ = x4 x3 ( 0 ) = 1
x4′ = −x1 − 2x3 + 1 x4 ( 0 ) = 0

In matrix form this becomes

⎡ x1′ ⎤ ⎡ 0 1 0 0⎤ ⎡ x1 ⎤ ⎡ 0 ⎤ ⎡ x1 ( 0 ) ⎤ ⎡0⎤
⎢ ′⎥ ⎢ ⎥ ⎢ ⎥ ⎢ −t ⎥ ⎢ ⎥ ⎢ ⎥
⎢ x2 ⎥⎢ 0 −3 −2 0 ⎥ ⎢ x2 ⎥ ⎢ e ⎥ ⎢ x2 ( 0 ) ⎥ ⎢1 ⎥
= + , = .
⎢ ′⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥ ⎢ x (0)⎥ ⎢1 ⎥
⎥ ⎢ ⎥
x3 0 0 0 1 x3 0 3 (0)
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥ ⎢ 0
′ 0 −2 0 ⎦ ⎣ x4 ⎦ x
⎣ x4 ⎦ ⎣ −1 1 4 ⎣ ⎦

34. y′′′ + y′ + 2z = 1, y ( 0 ) = 1, y ′ ( 0 ) = 0 , y ′′ ( 0 ) = 1

z ′ + y + 2z = sin t , z ( 0 ) = 1

Letting x1 = y , x2 = y ′ , x3 = y ′′ , x4 = z yields

x1′ = x2 x1 ( 0 ) = 0
x2′ = x3 x2 ( 0 ) = 0
.
x3′ = −x2 − 2x4 +1 x3 ( 0 ) = 1

x4′ = −x1 − 2x4 + sin t x4 ( 0 ) = 1


In matrix form this becomes

⎡ x1′ ⎤ ⎡ 0 1 0 0⎤ ⎡ x1 ⎤ ⎡ 0⎤ ⎡ x1 ( 0 ) ⎤ ⎡0⎤
⎢ ′⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥
444 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 444

⎢ x2 ⎥ = ⎢
In matrix form, this becomes 0 0 1 0 ⎥ ⎢ x2 ⎥ 0 ⎥ ⎢ x2 ( 0 ) ⎥ 0
+⎢ =⎢ ⎥.
⎢ x3′ ⎥ ⎢ 0 −1 0 −2⎥ ⎢ x3 ⎥ ⎢ 1⎥ ⎢ x3 ( 0 ) ⎥ ⎢1⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥

⎣ x4 ⎦ ⎣ −1 0 0 −2 ⎦ ⎣ x4 ⎦ ⎣sin t ⎦ x4 ( 0 ) ⎣1⎦
445 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 445

In matrix form, this becomes


 Conversion of Systems

35. x1 + x1 + 2x2 = e−t

x2 + 2x2 = 0
Letting z1 = x1 , z2 = x1 , z3 = x2 , z4 = x2 yields the system
z1 = z2
z2 = −z1 − 2z3 + e−t
z3 = z 4
z4 = −2z3 .
In matrix form this becomes
⎡ z1′ ⎤ ⎡ 0 1 0 0 ⎤ ⎡ z1 ⎤ ⎡ 0⎤
⎢ ′⎥ ⎢ ⎥ ⎢ ⎥ ⎢ −t ⎥
⎢ z2 ⎥ ⎢ −1 0 −2 0 ⎥ ⎢ z2 ⎥ ⎢ e ⎥
= + .
⎢ ′⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥
z3 0 0 0 1 z3 0
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥

⎣ 4 ⎦ ⎣ 0 0 −2 0 ⎦ ⎣ z4 ⎦
z 0

36. y ′′′ = f ( t, y, y′, y ′′, z, z ′ )


z ′′ = f ( t, y, y′, y ′′, z, z ′ )

Letting x1 = y , x2 = y′ , x3 = y ′ , x4 = z , x5 = z′ yields
x1′ = x2
x2′ = x3

x3′ = f ( t, x1 , x2 , x3 , x4 , x5 )
x4′ = x5
x5′ = g ( t, x1 , x2 , x3 , x4 , x5 ) .

37. x1 = a11 x1 + a12 x2 + a13 x3


x2 = a21 x1 + a22 x2 + a23 x3
x3 = a31 x1 + a32 x2 + a33 x3

If we let z1 = x1 , z2 = x1 , z3 = x2 , z4 = x2 , z5 = x3 , z6 = x3 , we get
z1 = z2

z2 = a11 z1 + a12 z3 + a13 z5


z3 = z 4
z4 = a21 z1 + a22 z3 + a23 z5
z5 = z 6
z6 = a31 z1 + a32 z3 + a33 z5 .
444 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 444

In matrix form z = Az , this becomes


⎡ z1 ⎤ ⎡ 0 1 0 0 0 0⎤ ⎡ z1 ⎤
⎢z ⎥ ⎢a 0 a12 0 a13 0 ⎥⎥ ⎢⎢ z2 ⎥⎥
⎢ 2⎥ ⎢ 11

⎢ z3 ⎥ ⎢ 0 0 0 1 0 0 ⎥ ⎢ z3 ⎥
⎢ ⎥ =⎢ ⎥ ⎢ ⎥.
⎢ z 4 ⎥ ⎢ a21 0 a22 0 a23 0 ⎥ ⎢ z4 ⎥

⎢ z5 ⎥ ⎢ 0 0 0 0 0 1 ⎥ ⎢ z5 ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
z6 ⎢⎣ a31 0 a32 0 a33 0 z6

 Solving Linear Systems


38. x1′ = x2
x2′ = −2x1 − 3x2

From first DE: x2 = x1′ . Substituting in second DE gives

( x1′ )′ = −2x1 − 3 ( x1′ )


or the second order DE
x1′ + 3x1′ + 2x1 = 0 .

Solving the second order DE gives


x1 = c1e −2t + c2 e−t .
Substituting this result in first DE gives
x2 = x1′ = −2c1e−2t − c2 e −t .

39. x1′ = 3x1 − 2x2


x2′ = 2x1 − 2x2

1 3
From first DE: x2 = − x′1 + x1 . Substituting in second DE yields a second order DE to solve for
2 2
x1 .

⎛ 1 3 ⎞′ ⎛ 1 3 ⎞
− x′ + x = 2x − 2 − x′ + x
⎜ 2 1 2 1⎟ 1 ⎜ 2 1 2 1⎟
⎝ ⎠ ⎝ ⎠
1 3
− x′′ + x′ = 2x + x′ − 3x
1
2 2 1 1 1 1

x1′′ − x1′ − 2x1 = 0


x1 = c1e 2t + c2 e −t

To find x2 , substitute the solution for x1 back into the first DE.

x =−
1
x′ +
3
x =−
1
( 2c e 2t
) 3 (c e
− c e −t + 2t
+ c e −t = ) 1
c e 2t + 2c e −t .
445 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 445
2 1 1 1 2 1 2 1 2
2 2 2 2 2
446 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 446

40. x1′ = x1 + x2
x2′ = 4x1 + x2

From first DE: x2 = x1′ − x1 . Substituting in second DE yields a second order DE to solve for x1 .

( x1′ − x1 )′ = 4x1 + ( x1′ − x1 )


x1′′ − x1′ = 4x1 + x1′ − x1
x1′′ − 2x1′ − 3x1 = 0

x1 = c1e3t + c2 e −t

( ) ( )
From first calculation, x2 = x1′ − x1 , so x2 = 3c1e3t − c2 e −t − c1e3t + c2 e −t = 2c1e3t − 2c2 e−t .

41. x1′ = x2 + t
x2′ = −2x1 + 3x2 + 5

From first DE: x2 = x1′ − t . Substituting in second DE yields a second order DE to solve for x1 .

( x1′ − t )′ = −2x1 + 3 ( x1′ − t ) + 5


x1′′ −1 = −2x1 + 3x1′ − 3t + 5

x1′′ − 3x1′ + 2x1 = −3t + 6.


x = c e 2t + c et
1h 1 2

To find x1 p by the method of undetermined coefficients, substitute x1 p = at + b , x1′ p = a , x1′p = 0

to obtain

0 − 3a + 2at + 2b = −3t + 6 .
Comparing like terms,
3
Coefficients of t: 2a = −3 so a = − .
2
3 3 3
Constants: −3a + 2b = 6 so b = . Hence x1 p = − t + . Therefore,
4 2 4

3 3
x1 = c1e 2t + c2et − t + .
2 4
From first calculation x2 = x1′ − t , so

3
x2 = 2c1e 2t + c2 et − −t .
2
447 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 447

 Solving IVPs for Systems


42. x1′ = 6x1 − 3x2
x2′ = 2x1 + x2

1
From first DE: x2 = 2x1 − x′1 . Substituting in second DE yields a second order DE to solve for x1 .
3

⎛ 1 ⎞′ ⎛ 1 ⎞
2x − x′ = 2x + 2x − x′
⎜ 1 3 1⎟ 1 ⎜ 1 3 1⎟
⎝ ⎠ ⎝ ⎠
x1′′ − 7x1′ +12x1 = 0
x1 = c1e3t + c2 e 4t .

1
From first calculation, x2 = 2x1 − x′1 , so
3

(
x2 = 2 c1e3t + c2 e4t − ) 31 (3c e1
3t
)
+ 4c2 e 4t = c1 e3t +
2 4t
3
c2 e .

Applying initial conditions:


x1 ( 0 ) = 2 ⇒ c1 + c2 = 2

2
x 0 =3 ⇒ c + c =3
2( ) 1 2
3

so
c2 = −3 and c1 = 5 .

The solution to the IVP is x1 = 5e3t − 3e 4t , x2 = 5e3t − 2e 4t .


43. x1′ = 3x1 + 4x2
x2′ = 2x1 + x2

1 3
From first DE: x2 = x′1 − x1 . Substituting in second DE yields a second order DE to solve for
4 4
x1 .

⎛1 3 ⎞′ ⎛1 3 ⎞
x′ − x = 2x + x′ − x
⎜ 4 1 4 1⎟ 1 ⎜ 4 1 4 1⎟
⎝ ⎠ ⎝ ⎠
x1′′ − 3x1′ = 8x1 + x1′ − 3x1
x1′′ − 4x1′ − 5x1 = 0
x1 = c1e5t + c2 e −t

1 3
From first calculation, x2 = x′1 − x1 , so
4 4
448 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 448

x2 =
1
4
( ) (3
4
) 1
5c1e5t − c2 e −t − c1 e5t + c2 e −t = c1 e5t − c2 e −t
2
449 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 449

Applying initial conditions:


x1 ( 0 ) = 1 ⇒ c1 + c2 = 1

1
x 0 = −1 ⇒ c − c = −1
2( ) 1 2
2

so
c1 = 0 and c2 = 1.

The solution to the IVP is x1 = e−t , x2 = −e−t .

 Counterexample
44. An example: The degenerate system
x1 + x2 + x1 = 0
x1 + x2 + x1 = 0

where both equations are exactly the same clearly cannot be written as a second-order equation in
either x1 or x2 . The reader might contemplate finding all the solutions of such an undetermined

system.

Another approach: Note that when we write an nth-order equation such as


ay ′ + by′ + cy = 0
as a system of first-order equations by letting x1 = y , x2 = y ′ , the system has the form

⎡ 0 1⎤
⎡x ⎤ ⎡x ⎤
⎢ b⎥
⎢x ⎥ =⎢ − c
1 1
.
− ⎥⎢ x⎥
⎣ 2⎦ ⎣ 2⎦
a a

This shows we cannot obtain a second-order equation in x1 with x2 = x1 unless the coefficient

matrix has the preceding form in which the first row contains a 0 and 1. Hence, a system such as
⎡ x1 ⎤ ⎡1 1⎤ ⎡ x1 ⎤
⎢ x ⎥ = ⎢ 4 1⎥ ⎢ x ⎥
⎣ 2⎦ ⎣ ⎦ ⎣ 2⎦

cannot be transformed into a second-order equation in x1 with x2 = x1 .

 Coupled Mass-Spring System


45. Given the linear system
mx1 = −k1 x1 + k2 ( x2 − x1 ) = − ( k1 + k 2 ) x1 + k 2 x2
450 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 450

mx2 = −k2 ( x2 − x1 ) = k2 x1 − k2 x2 ,
451 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 451

we let
z1 = x1 z 3 = x2
z2 = x1 z4 = x2 .

We then have the first-order system


z1 = z2
k1 +k 2 k2
z z z
2 =− 1 + 3
m m
z3 = z4

⎛k ⎞ ⎛k ⎞
z4 = ⎜ 2 ⎟ z1 − ⎜ 2 ⎟ z3 .
⎝m⎠ ⎝m⎠
In matrix form this becomes
⎡ 0 1 0 0⎤
⎡z ⎤ ⎢ ⎥ ⎡z ⎤
1
⎢ ⎥ ⎢ − ( k1 +k2 ) 0
k2 ⎥
0 ⎢ ⎥
1

⎢ z2 ⎥ = ⎢ m m ⎥ ⎢ z2 ⎥
.
⎢ z3 ⎥ ⎢ ⎥
0 0 0 1 ⎢ z3 ⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥
⎣ z4 ⎦ ⎢ k2 k2 ⎥ ⎣ z4 ⎦
0 − 0
⎢ ⎥
⎣ m m ⎦

 Satellite Problem
k
46. r = r ( t )θ 2 ( t ) − + u1 ( t )
r 2 (t )

2θ ( t ) r ( t ) 1
()
r (t ) r (t )
2

θ= + u t
Letting
x1 = r x3 = θ
x2 = r x4 = θ ,

we have the system


x1 = x2
k
x2 = x1 x42 − + u1 ( t )
x12
x3 = x4
2 x2 x4 1
x4 = + u2 ( t ) .
452 CHAPTER 4 Higher-Order Linear Differential Equations SECTION 4.7 Conservation and Conversion 452

x1 x1
SECTION 4.7 Conservation and Conversion 449

 Two Inverted Pendulums

47. θ1 = ( mg +1)θ1 + mgθ 2 − u ( t )


θ2 = mgθ1 + ( mg +1)θ 2 − u ( t )

Letting
x1 = θ1
x2 = 1θ
x3 = θ 2
x4 = θ 2

we have first-order linear system


x1 = x2
x2 = ( mg +1) x1 + mgx3 − u ( t )
x3 = x4
x4 = mgx1 + ( mg +1) x3 − u ( t ) .

In matrix form this becomes


⎡ x1 ⎤ ⎡ 0 1 0 0 ⎤ ⎡ x1 ⎤ ⎡ 0⎤
⎢x ⎥ ⎢ ⎢ ⎥
⎢ 2 ⎥ = ⎢ mg +1 0 mg 0⎥⎥ ⎢ x2 ⎥ ⎢⎢ −u(t) ⎥⎥
+ .

⎢ x3 ⎥ ⎢ 0 0 0 1 ⎥ ⎢ x3 ⎥ ⎢ 0⎥
⎢ ⎥ ⎢ ⎥⎢ ⎥ ⎢ ⎥
⎣ x4 ⎦ ⎣ mg 0 mg +1 0 ⎦ ⎣ x4 ⎦ ⎣ −u(t) ⎦

 Suggested Journal Entry

48. Student Project

You might also like