Chapter 5

The Laplace Transform

5.1 Definition of the Laplace Transform

1.

The function is piecewise continuous.

2.

The function is neither continuous nor piecewise continuous.

285
286 CHAPTER 5. THE LAPLACE TRANSFORM

3.

The function is piecewise continuous.

4.

The function is piecewise continuous.

5. The function is of exponential order. We can take K = 2, a = 3 and M = 0.
6. Since f (t) = −5e−6t implies |f (t)| = 5e−6t , the function is of exponential order. We can
take K = 5, a = −6 and M = 0.
7. Since |f (t)| = |e5t sin(3t)| ≤ e5t , the function is of exponential order. We can take K = 1,
a = 5 and M = 0.
8. Since t = t1 = eln t ≤ et for all t ≥ 0, the function is of exponential order. We can take
K = 1, a = 1 and M = 0.
2 2
9. cosh t2 = (et + e−t )/2 is not of exponential order, since for all a > 0, there exists an M
2
such that et ≥ eat for all t ≥ M .
10. Since | sin(t4 )| ≤ 1 = e0·t , the function is of exponential order. We can take K = 1,
a = 0 and M = 0.
11. Since 1/(1 + t) ≤ 1 for all t ≥ 0, the function is of exponential order. We can take
K = 1, a = 0 and M = 0.
3
12. et is not of exponential order, since for all a > 0, there exists an M > 0 such that
3
et ≥ eat for all t ≥ M .
5.1. DEFINITION OF THE LAPLACE TRANSFORM 287

13.(a) For f (t) = t, the Laplace transform is

Z ∞ ∞
−st −e−st t − e−st  1
F (s) = e t dt = 2
= 2.
0 s 
t=0 s

(b) For f (t) = t2 , the Laplace transform is

Z ∞ ∞
−st 2 e−st s2 t2 + 2e−st st + 2e−st  2
F (s) = e t dt = − = .
0 s3 
t=0 s3

(c) For f (t) = tn , using induction, the Laplace transform is

Z ∞
n!
F (s) = e−st tn dt = n+1 .
0 s

14. For 
t 0≤t≤1
f (t) =
6 1<t
the Laplace transform
Z ∞ Z 1 Z ∞
−st −st
F (s) = e f (t) dt = e t dt + 6e−st dt.
0 0 1

We integrate the first integral on the right-hand side by parts. In particular, we have
Z t 1 Z 1 −st
te−st  e−s 1 1 −st
Z
−st e
e t dt = − dt = + e dt
0 −s 0 0 −s −s s 0
1
e−s 1 e−st  e−s e−s 1
= + · = − 2 + 2.
−s s −s  −s s s 0

The second integral above can be evaluated as follows:

Z ∞ ∞
−st e−st  e−s
6e dt = 6 = 6 .
1 −s 1 s
Combining these two integrals, we have
e−s e−s 1 e−s 1 e−s e−s
F (s) = − 2 + 2 +6 = 2 − 2 +5 .
−s s s s s s s
15. For 
 0 0≤t≤1
f (t) = 1 1<t≤2
0 2 < t,


the Laplace transform is given by

Z 1 2 ∞ 2
2
e−st 
Z Z Z
−st −st −st −st
F (s) = e · 0 dt + e dt + e · 0 dt = e dt =
0 1 2 1 −s 1
−2s
e e−s e −s
− e−2s
= − = .
−s −s s
288 CHAPTER 5. THE LAPLACE TRANSFORM

16. For 
0 0≤t≤1
f (t) =
e−t 1<t
the Laplace transform is given by
Z ∞ Z ∞ ∞
−st −t −(s+1)t e−(s+1)t  e−(s+1) e−(s+1)
F (s) = e e dt = e dt = = 0 − = .
1 1 −(s + 1) 1 −(s + 1) s+1

17. For  2
 t 0≤t≤1
f (t) = 5−t 1<t≤2
6 2<t


the Laplace transform is given by

Z 1 Z 2 Z ∞
−st 2 −st
F (s) = e t dt + e (5 − t) dt + 6e−st dt.
0 1 2

For the first integral on the right-hand side, we proceed as follows:

Z 1 1 Z 1 −st
t2 e−st  e−s 2 1 −st
Z
−st 2 e
e t dt = − · 2t dt = + te dt.
0 −s 0 0 −s −s s 0

By the solution to problem 14, we know that

Z 1
−st e−s e−s 1
te dt = − 2 + 2.
0 −s s s
Therefore,
1
e−s 2 e−s e−s e−s 2e−s 2e−s
Z  
−st 2 1 2
e t dt = + − 2 + 2 = − 2 − 3 + 3.
0 −s s −s s s −s s s s
For the second integral on the right-hand side, we begin by separating it into two pieces.
Z 2 Z 2 Z 2
−st −st
e (5 − t) dt = 5 e dt − te−st dt ≡ A + B.
1 1 1

Now, for A, we have 2

2
5e−st  5e−s 5e−2s
Z
−st
5 e dt = = − .
1 −s 1 s s
For term B, we have
" Z 2 −st #
2 −st 2
e−s 1 2 −st
Z   −2s Z 
−st te e 2e
− te dt = −  − dt = − − + e dt
1 −s 1 1 −s −s −s s 1
" 2 #
2e−2s e−s 1 e−st 
 −2s
e−s e−2s e−s

2e
= − + = · =− + − 2 + 2 .
−s s s −s 1 −s s s s
5.1. DEFINITION OF THE LAPLACE TRANSFORM 289

Now combining the results for A and B, we have

Z 2
5e−s 5e−2s
 −2s
e−s e−2s e−s

−st 2e
e (5 − t) dt = − − + − 2 + 2 .
1 s s −s s s s
Finally, for the third integral above, we have
Z ∞ −st ∞ −2s
6e−2s

6e  = 0 − 6e
6e−st dt = = .
2 −s 2 −s s
Combining these integrals, we have
e−s 2e−s 2e−s 2 5e−s 5e−2s 2e−2s e−s e−2s e−s 6e−2s
F (s) = − 2 − 3 + 3+ − + − + 2 − 2 + .
−s s s s s s s s s s s
Simplifying, we conclude that the Laplace transform is given by
3e−s 3e−s 3e−2s e−2s 2e−s 2
F (s) = − 2 + + 2 − 3 + 3.
s s s s s s

18. The Laplace transform is given by

Z a
1 a (b−s)t 1 a −(b+s)t 1 1 − e(b−s)a 1 1 − e−(b+s)a
Z Z    
−st
(cosh bt)e dt = e dt + e dt = + .
0 2 0 2 0 2 s−b 2 s+b
Taking the limit as a → ∞, we conclude that
Z ∞    
−st 1 1 1 1 s
(cosh bt)e dt = + = 2 .
0 2 s−b 2 s+b s − b2
This function is valid as long as s > |b|.
19. The Laplace transform is given by
Z a
1 a (b−s)t 1 a −(b+s)t 1 1 − e(b−s)a 1 1 − e−(b+s)a
Z Z    
−st
(sinh bt)e dt = e dt − e dt = − .
0 2 0 2 0 2 s−b 2 s+b
Taking the limit as a → ∞, we conclude that
Z ∞    
−st 1 1 1 1 b
(sinh bt)e dt = − = 2 .
0 2 s−b 2 s+b s − b2
This function is valid as long as s > |b|.
20. We note that eat cosh bt = (e(a+b)t + e(a−b)t )/2. Thefore,
Z A
1 1 − e(a+b−s)A 1 1 − e−(b−a+s)A
   
at −st
(e cosh bt)e dt = + .
0 2 s−a+b 2 s+b−a
Taking the limit as A → ∞,
Z ∞    
at −st 1 1 1 1 s−a
(e cosh bt)e dt = + = .
0 2 s−a+b 2 s+b−a (s − a)2 − b2
290 CHAPTER 5. THE LAPLACE TRANSFORM

This limit exists as long as s − a > |b|.

21. We note that eat sinh bt = (e(a+b)t − e(a−b)t )/2. Therefore,
Z A
1 1 − e(a+b−s)A 1 1 − e−(b−a+s)A
   
at −st
(e sinh bt)e dt = − .
0 2 s−a+b 2 s+b−a
Taking the limit as A → ∞,
Z ∞    
at −st 1 1 1 1 b
(e sinh bt)e dt = − = .
0 2 s−a+b 2 s+b−a (s − a)2 − b2

This limit exists as long as s − a > |b|.

22. Since the Laplace transform is a linear operator, we have
1 1
L[cos bt] = L[eibt ] + L[e−ibt ].
2 2
Now Z ∞
1
e±ibt e−st dt = .
0 s ∓ ib
Therefore,  
1 1 1 s
L[cos bt] = + = 2 .
2 s − ib s + ib s + b2
23. Since the Laplace transform is a linear operator, we have
1 1
L[eat sin bt] = L[e(a+ib)t ] − L[e(a−ib)t ].
2i 2i
Now Z ∞
1
e(a±ib)t e−st dt = .
0 s − a ∓ ib
Therefore (for s > a),
 
at 1 1 1 b
L[e sin bt] = − = .
2i s − a − ib s − a + ib (s − a)2 + b2

24. Since the Laplace transform is a linear operator, we have

1 1
L[eat cos(bt)] = L[e(a+ib)t ] + L[e(a−ib)t ].
2 2
Now Z ∞
1
e(a±ib)t e−st dt = .
0 s − a ∓ ib
Therefore (for s > a),
 
at 1 1 1 s−a
L[e cos(bt)] = + = .
2 s − a − ib s − a + ib (s − a)2 + b2
5.1. DEFINITION OF THE LAPLACE TRANSFORM 291

25. Integrating by parts, we have

A
A Z A
te(a−s)t  1 − eA(a−s) + A(a − s)eA(a−s)
Z
at −st 1 (a−s)t
te · e dt = − + e dt = .
0 s − a 0 0 s−a (s − a)2

Taking the limit as A → ∞ (when s > a),

Z ∞
1
eat · e−st dt = .
0 (s − a)2

26. We will use the fact that sin at = (eiat − e−iat )/2i. Therefore,
Z A Z A Z A 
−st 1 (ia−s)t −(ia+s)t
(t sin at)e dt = te dt − te dt
0 2i 0 0
" # " #
(ia−s)t A Z A (−ia−s)t A Z A
 
1 te 1 1 te 1
= −  + e(ia−s)t dt − −  + e(−ia−s)t dt
2i s − ia 0 
0 s − ia 2i s + ia 0 
0 s + ia

Ae(ia−s)A Ae(−ia−s)A
 
1 1 1
= − + + 2
− .
2i s − ia s + ia (s − ia) (s + ia)2

Taking the limit as A → ∞, we have

Z ∞
1 (s + ia)2 − (s − ia)2
 
−st 2as
(t sin at)e dt = 2 2
= 2 .
0 2i (s − ia) (s + ia) (s + a2 )2

27. We will use the fact that t cosh at = (teat + te−at )/2. We note that for any value of c,
A
A Z A (c−s)t
te(c−s)t  1 − eA(c−s) + A(c − s)eA(c−s)
Z
ct −st e
te · e dt = − + dt = .
0 s − c 0 0 s−c (s − c)2

Taking the limit as A → ∞, we have

Z ∞
1
tect · e−st dt = .
0 (s − c)2

Therefore (for s > |a|),

Z ∞
s2 + a2
 
−st 1 1 1
(t cosh at) · e dt = + = .
0 2 (s − a)2 (s + a)2 (s − a)2 (s + a)2

28. Integrating by parts, we see that

A
A Z A
tn e(a−s)t 
Z
n at −st n n−1 (a−s)t
t e ·e dt = − + t e dt
0 s−a 0 
0 s−a
Z A
An e−(s−a)A n n−1 (a−s)t
= − + t e dt.
s−a 0 s−a
292 CHAPTER 5. THE LAPLACE TRANSFORM

Continuing to integrate by parts, we conclude that

Z A
An e(a−s)A nAn−1 e(a−s)A n!Ae(a−s)A n!(e(a−s)A − 1)
tn eat · e−st dt = − − − − . . . − .
0 s−a (s − a)2 (n − 2)!(s − a)3 (s − a)n+1
We notice that as A → ∞, all of the terms above except the last one go to zero. We conclude
that Z ∞
n!
tn eat · e−st dt = .
0 (s − a)n+1

29. We will use the fact that sin at = (eiat − e−iat )/2i. Therefore,
Z A Z A Z A 
2 −st 1 2 (ia−s)t 2 −(ia+s)t
(t sin at)e dt = te dt − te dt .
0 2i 0 0

Now for any value of c, we note that

Z A A
2 (c−s)t e(c−s)t (c − s)2 t2 − 2e(c−s)t (c − s)t + 2e(c−s)t 
te dt =
0 (c − s)3 
0
e(c−s)A (c − s)2 A2 − 2e(c−s)A (c − s)A + 2e(c−s)A − 2
= .
(c − s)3
Taking the limit as A → ∞, we have
Z ∞
2
t2 e(c−s)t dt = − .
0 (c − s)3
Therefore (for s > 0),
Z ∞
2a(3s2 − a2 )
 
2 −st 1 2 2
(t sin at)e dt = − + = .
0 2i (ia − s)3 (ia + s)3 (s2 + a2 )3

30. We will use the fact that t2 sinh at = (t2 eat − t2 e−at )/2. Using the fact that for any c,
Z ∞
2
t2 e(c−s)t dt = − ,
0 (c − s)3
we see that
∞
2a(3s2 + a2 )
Z  
2 −st 1 2 2
(t sinh at)e dt = − = .
0 2 (s − a)3 (s + a)3 (s2 − a2 )3

31. First, we note that

Z A
(t2 + 1)−1 dt = arctan t|A
0 = arctan A − arctan 0 = arctan A.
0

Therefore, as A → ∞, we have
Z ∞
π
(t2 + 1)−1 dt = lim arctan A = .
0 A→∞ 2
5.1. DEFINITION OF THE LAPLACE TRANSFORM 293

Therefore, the integral converges.

32. Integrating by parts, we see that
Z A
te−t dt = 1 − e−A − Ae−A .
0

As A → ∞, 1 − e−A − Ae−A → 1. Therefore, the integral converges.

33. Based on a series expansion, we see that for t > 0,
t2 t3 t4 t4
et > 1 + t + + + > .
2 6 24 24
Therefore, t−4 et > 1/24. Thus
A A
A−1
Z Z
−4 t 1
t e dt > dt = .
1 1 24 24
As A → ∞, we see that (A − 1)/24 → ∞. Therefore, the integral diverges.
34. We can see that
∞ ∞ ∞
Z  Z Z
−t −t
e−t dt = 1.
 

 e cos t dt ≤ |e cos t| dt ≤
0 0 0

Therefore, the integral converges.

35.(a) We see that
M +1 ∞
Z Z 
e−st f (t) dt + e−st f (t) dt
 
|F (s)| = 
0 M +1
Z M +1 Z ∞
−st
≤ max |f (t)| e dt + K e−st eat dt
0≤t≤M +1 0 M +1
−st M +1 −(s−a) ∞
 
e  e t
= max |f (t)| +K
0≤t≤M +1 −s 0 −(s − a) M +1
1 − e−(M +1)t K −(s−a)(M +1)
= max |f (t)| + e .
0≤t≤M +1 s s−a

(b) Since e−(s−a)(M +1) → 0 as s → ∞ and (s − a)/s → 1 as s → ∞, there exists a constant

K1 such that Ke−(s−a)(M +1) < K1 (s − a)/s for s sufficiently large.
(c) Using the results from parts (a) and (b) and the fact that (1 − e−(M +1)t )/s < 1/s, we
conclude that
1 K1
|F (s)| ≤ max |f (t)| +
0≤t≤M +1 s s
for s sufficiently large.
36.(a) Integrating by parts,
Z A Z A Z A
−x p −x p A −x p−1 p −A
e−x xp−1 dx.

e x dx = −e x 0 + p e x dx = −A e +p
0 0 0
294 CHAPTER 5. THE LAPLACE TRANSFORM

Taking the limit as A → ∞, we see that

Z ∞ Z ∞
−x p
Γ(p + 1) = e x dx = p e−x xp−1 dx = pΓ(p).
0 0

(b) We see that Z ∞

Γ(1) = e−x dx = 1.
0

(c) Let p = n. Using the result in part (a), we see that

Γ(n + 1) = nΓ(n) = n(n − 1)Γ(n − 1) = . . . = n(n − 1)(n − 2) · 2 · Γ(1).
Since Γ(1) = 1, we conclude that Γ(n + 1) = n!.
(d) Using the result in part (a), we see that
Γ(p + n) = (p + n − 1)Γ(p + n − 1) = (p + n − 1)(p + n − 2)Γ(p + n − 2)
= (p + n − 1)(p + n − 2) · · · (p + 1)pΓ(p).
Therefore,
Γ(p + n)
= p(p + 1)(p + 2) · · · (p + n − 1).
Γ(p)
√
Given that Γ(1/2) = π, it follows that
    √
3 1 1 π
Γ = Γ =
2 2 2 2
and     √
11 9 7 5 3 3 945 π
Γ = · · · Γ = .
2 2 2 2 2 2 32
37.(a) By definition, Z ∞
p
L(t ) = e−st tp dt.
0
Now letting x = st, we see that dx = s dt. Therefore,
Z ∞ Z ∞ Z ∞
−st x
 p dx 1 Γ(p + 1)
−st p
p
L(t ) = e t dt = e . = p+1 e−st xp dx = .
0 0 s s s 0 sp+1

(b) Using part (a) and the fact that for n a positive integer, Γ(n + 1) = n!, we conclude that
L(tn ) = n!/sn+1 .
(c) By definition, Z ∞
−1/2
L(t )= e−st t−1/2 dt.
0
√
Making the change of variables x = st, we see that x2 = st implies 2x dx = s dt. Therefore,
Z ∞ Z ∞  2 −1/2 Z ∞
−st −1/2 −x2 x 2x dx 2 2
e t dt = e =√ e−x dx.
0 0 s s s 0
5.2. PROPERTIES OF THE LAPLACE TRANSFORM 295

(d) Using the result from part (a), we see that

Γ(3/2)
L(t1/2 ) = .
s3/2
√
Then, using the result from problem 36, we know that Γ(3/2) = π/2. Therefore,
√
1/2 π
L(t ) = 3/2 .
2s

5.2 Properties of the Laplace Transform

3
1. By Example 7 in Section 5.1, for f (t) = sin 3t, F (s) = 2 . Therefore, by Theorem
s +9
5.2.1,
3
L(e−2t sin 3t) = F (s + 2) = .
(s + 2)2 + 9

2. First, we compute the Laplace transform of f (t) = cos 2t. For f (t) = cos 2t,
Z ∞
1 ∞ −st 2it
Z
−st
F (s) = e cos 2t dt = e (e + e−2it ) dt
0 2 0
" #
Z ∞ −(s−2i)t −(s+2i)t b
1 1 e e
= (e−(s−2i)t + e−(s+2i)t ) dt = lim + 
2 0 b→∞ 2 −(s − 2i) −(s + 2i) 0
 
1 1 1 2
= + = 2 .
2 s − 2i s + 2i s +4

Therefore, by Theorem 5.2.1,

s−3
L(e3t cos 2t) = F (s − 3) = .
(s − 3)2 + 4

n!
3. By Corollary 5.2.5, L(tn ) = . Combining this fact with linearity, we have
sn+1
6! 2! 0! 720 8 5
L(t6 − 4t2 + 5) = L(t6 ) − 4L(t2 ) + 5L(1) = 7
−4· 3 +5· 1 = 7 − 3 + .
s s s s s s
s
4. By Theorem 5.2.4, L(t cos 3t) = (−1)F 0 (s) where F (s) = L(cos 3t) = . Then
s2 +9
9 − s2
 
0 d s
F (s) = = .
ds s2 + 9 (s2 + 9)2

Therefore,
s2 − 9
L(t cos 3t) = .
(s2 + 9)2
296 CHAPTER 5. THE LAPLACE TRANSFORM

5. By Theorem 5.2.1, L(e−2t (t3 + 1)2 ) = F (s + 2) where F (s) = L((t3 + 1)2 ). Using linearity
and Corollary 5.2.5, we have
6! 12 1
L((t3 + 1)2 ) = L(t6 + 2t3 + 1) = + 4 + .
s7 s s
Therefore,
6! 12 1
L(e−2t (t3 + 1)2 ) = + + .
(s + 2)7 (s + 2)4 s + 2

6. By Theorem 5.2.1, L(t6 e5t ) = F (s − 5) where F (s) = L(t6 ). By Corollary 5.2.5, we know
6!
that L(t6 ) = 7 . Therefore,
s
6!
L(t6 e5t ) = .
(s − 5)7

7. By Theorem 5.2.4, L(t2 sin bt) = F 00 (s) where F (s) = L(sin bt). By example 7 of Section
5.1,
b
F (s) = L(sin bt) = 2 .
s + b2
Differentiating, we get

−2sb 6bs2 − 2b3

F 0 (s) = and F 00 (s) = .
(s + b2 )2
2 (s2 + b2 )3

Therefore,
6bs2 − 2b3
L(t2 sin bt) = .
(s2 + b2 )3

n!
8. By Theorem 5.2.1, L(eat tn ) = F (s − a) where F (s) = L(tn ) = . Therefore,
sn+1
n!
L(eat tn ) = .
(s − a)n+1

9. By Theorem 5.2.4, L(teat sin bt) = (−1)F 0 (s) where F (s) = L(eat sin bt). Further,
b
L(eat sin bt) = G(s − a) where G(s) = L(sin bt) = 2 . Therefore,
s + b2
b
L(eat sin(bt)) =
(s − a)2 + b2

and  
d b −2b(s − a)
2 2
= .
ds (s − a) + b ((s − a)2 + b2 )2
Therefore,
2b(s − a)
L(teat sin bt) = .
((s − a)2 + b2 )2
5.2. PROPERTIES OF THE LAPLACE TRANSFORM 297

10. By Theorem 5.2.4, L(teat cos bt) = (−1)F 0 (s) where F (s) = L(eat cos bt). Further,
s
L(eat cos bt) = G(s − a) where G(s) = L(cos bt) = 2 . Now
s + b2
s−a
L(eat cos bt) =
(s − a)2 + b2
and
b2 − (s − a)2
 
d s−a
= .
ds (s − a)2 + b2 ((s − a)2 + b2 )2
Therefore,
(s − a)2 − b2
L(teat cos bt) = .
((s − a)2 + b2 )2
Z t
11.(a) Let g1 (t) = f (t1 ) dt1 . Therefore, g10 (t) = f (t) which implies L(g10 (t)) = L(f (t)) =
0
F (s). Also, by Theorem 5.2.2, L(g10 (t)) = sL(g1 (t)) − g1 (0) = sL(g1 (t)). Therefore,
sL(g1 (t)) = L(g10 (t)) = F (s), so
F (s)
L(g1 (t)) = .
s
That is, Z t 
1
L f (τ ) dτ = F (s).
0 s
(b) Let Z tZ tn Z t2
gn (t) = ··· f (t1 ) dt1 · · · dtn .
0 0 0
Therefore, Z tZ tn−1 Z t2
gn0 (t) = ··· f (t1 ) dt1 · · · dtn−1 .
0 0 0
(n) (n)
Continuing, we conclude that gn (t) = f (t). Therefore, L(gn (t)) = L(f (t)) = F (s). But
also, by Corollary 5.2.3,

L(gn(n) (t)) = sn L(gn (t)) − sn−1 gn (0) − . . . − sgn(n−2) (0) − gn(n−1) (0) = sn L(gn (t)).

Therefore,
1 1
L(gn (t)) = n
L(gn(n) (t)) = n F (s).
s s
12. The equation is y 00 + 2y 0 − 2y = 0; applying the Laplace transform we get L(y 00 ) +
2L(y 0 ) − 2L(y) = 0. Then [s2 L(y) − sy(0) − y 0 (0)] + 2[sL(y) − y(0)] − 2L(y) = 0, so
[s2 + 2s − 2]L(y) − 2s − 1 − 4 = 0. Thus [s2 + 2s − 2]L(y) − 2s − 5 = 0, and then
2s + 5
L(y) = .
s2 + 2s − 2

13. The equation is 9y 00 + 12y 0 + 4y = 0; applying the Laplace transform we get 9L(y 00 ) +
12L(y 0 ) + 4L(y) = 0. Then 9[s2 L(y) − sy(0) − y 0 (0)] + 12[sL(y) − y(0)] + 4L(y) = 0, so
298 CHAPTER 5. THE LAPLACE TRANSFORM

9[s2 L(y) − 2s + 1] + 12[sL(y) − 2] + 4L(y) = 0. Thus [9s2 + 12s + 4]L(y) − 18s + 9 − 24 = 0,

and then
18s + 15
L(y) = 2 .
9s + 12s + 4
14. The equation is y 00 + 3y 0 + 2y = 0; applying the Laplace transform we get L(y 00 ) +
3L(y 0 ) + 2L(y) = 0. Then [s2 L(y) − sy(0) − y 0 (0)] + 3[sL(y) − y(0)] + 2L(y) = 0, so
[s2 L(y) − 3s + 1] + 3[sL(y) − 3] + 2L(y) = 0. Thus [s2 + 3s + 2]L(y) − 3s + 1 − 9 = 0, and
then
3s + 8
L(y) = 2 .
s + 3s + 2
15. The equation is 6y 00 + 5y 0 + y = 0; applying the Laplace transform we get 6L(y 00 ) +
5L(y 0 ) + L(y) = 0. Then 6[s2 L(y) − sy(0) − y 0 (0)] + 5[sL(y) − y(0)] + L(y) = 0, so 6[s2 L(y) −
4s] + 5[sL(y) − 4] + L(y) = 0. Thus [6s2 + 5s + 1]L(y) − 24s − 20 = 0, and then
24s + 20
L(y) = .
6s2 + 5s + 1

16. The equation is y 00 − 2y 0 + 2y = t2 et + 7; applying the Laplace transform we get L(y 00 ) −

2L(y 0 )+2L(y) = L(t2 et +7), so [s2 L(y)−sy(0)−y 0 (0)]−2[sL(y)−y(0)]+2L(y) = L(t2 et +7).
Now L(t2 et + 7) = L(t2 et ) + 7L(1). First, L(t2 et ) = F (s − 1) where F (s) = L(t2 ) = 2/s3 .
Therefore, L(t2 et ) = 2/(s − 1)3 . Further, 7L(1) = 7/s. Thus
2 7
L(t2 et + 7) = 3
+ .
(s − 1) s
Therefore,
2 7
[s2 L(y) − sy(0) − y 0 (0)] − 2[sL(y) − y(0)] + 2L(y) = 3
+ ,
(s − 1) s
so
2 7
[s2 L(y) − s − 1] − 2[sL(y) − 1] + 2L(y) = + ,
(s − 1)3 s
and then
2 7
[s2 − 2s + 2]L(y) − s − 1 + 2 = 3
+ .
(s − 1) s
We obtain  
1 2 7
L(y) = 2 · + +s−1 .
(s − 2s + 2) (s − 1)3 s

17. The equation is y 00 − 5y 0 − 6y = t2 + 7; applying the Laplace transform we get L(y 00 ) −

5L(y 0 ) − 6L(y) = L(t2 + 7). This implies that
2! 7
[s2 L(y) − sy(0) − y 0 (0)] − 5[sL(y) − y(0)] − 6L(y) = + ,
s3 s
thus
2 7
[s2 L(y) − s] − 5[sL(y) − 1] − 6L(y) = 3
+ .
s s
5.2. PROPERTIES OF THE LAPLACE TRANSFORM 299

Simplifying, we get
2 7
[s2 − 5s − 6]L(y) − s + 5 =
3
+ ,
s s
and we obtain  
1 2 7
L(y) = 2 + +s−5 .
s − 5s − 6 s3 s

18. The equation is y 00 + 4y = 3e−2t sin 2t; applying the Laplace transform we get L(y 00 ) +
4L(y) = 3L(e−2t sin 2t). Thus [s2 L(y) − sy(0) − y 0 (0)] + 4L(y) = 3F (s + 2), where F (s) =
L(sin 2t) = 2/(s2 + 4). Therefore,
6
[s2 L(y) − sy(0) − y 0 (0)] + 4L(y) = ,
(s + 2)2 + 4
which means
6
[s2 L(y) − 2s + 1] + 4L(y) = .
(s + 2)2 + 4
Thus
6
[s2 + 4]L(y) − 2s + 1 = ,
(s + 2)2 + 4
and we obtain  
1 6
L(y) = 2 + 2s − 1 .
s + 4 (s + 2)2 + 4

19. The equation is y 00 + 2y 0 + 5y = t cos 2t; applying the Laplace transform we get L(y 00 ) +
2L(y 0 ) + 5L(y) = L(t cos 2t). Thus [s2 L(y) − sy(0) − y 0 (0)] + 2[sL(y) − y(0)] + 5L(y) =
(−1)F 0 (s), where F (s) = L(cos 2t) = s/(s2 + 4). Therefore, F 0 (s) = (4 − s2 )/(s2 + 4)2
implies
s2 − 4
L(t cos 2t) = 2 .
(s + 4)2
We obtain
s2 − 4
[s2 L(y) − s] + 2[sL(y) − 1] + 5L(y) = ,
(s2 + 4)2
which implies that
s2 − 4
[s2 + 2s + 5]L(y) − s − 2 = .
(s2 + 4)2
Then  2 
1 s −4
L(y) = 2 +s+2 .
s + 2s + 5 (s2 + 4)2

20. The equation is y 000 + y 00 + y 0 + y = 0; applying the Laplace transform we get L(y 000 ) +
L(y 00 ) + L(y 0 ) + L(y) = 0. Thus [s3 L(y) − s2 y(0) − sy 0 (0) − y 00 (0)] + [s2 L(y) − sy(0) − y 0 (0)] +
[sL(y)−y(0)]+L(y) = 0, which implies [s3 L(y)−4s2 +2]+[s2 L(y)−4s]+[sL(y)−4]+L(y) = 0.
After simplification, [s3 + s2 + s + 1]L(y) − 4s2 + 2 − 4s − 4 = 0, so we obtain that

4s2 + 4s + 2
L(y) = .
s3 + s2 + s + 1
300 CHAPTER 5. THE LAPLACE TRANSFORM

21. The equation is y 0000 − 6y = te−t ; applying the Laplace transform we get L(y 0000 ) − 6L(y) =
L(te−t ). Thus [s4 L(y) − s3 y(0) − s2 y 0 (0) − sy 00 (0) − y 000 (0)] − 6L(y) = F (s + 1), where
F (s) = L(t) = 1/s2 . Therefore, L(te−t ) = 1/(s + 1)2 , which implies
1
[s4 L(y) − 9] − 6L(y) = .
(s + 1)2
Thus
1
[s4 − 6]L(y) − 9 = ,
(s + 1)2
and then  
1 1
L(y) = 4 +9 .
s − 6 (s + 1)2
1
22. The equation is Lq 00 + Rq 0 + q = e(t); applying the Laplace transform we get LL(q 00 ) +
c
1 1
RL(q ) + L(q) = L(e). Thus L[s2 L(q) − sq(0) − q 0 (0)] + R[sL(q) − q(0)] + L(q) = E(s),
0
c  c
2 1 0
which is Ls + Rs + L(q) − Lsq(0) − Lq (0) − Rq(0) = E(s). This implies
c
Lsq(0) + Lq 0 (0) + Rq(0) + E(s) (Ls + R)q(0) + Li(0) + E(s)
Q(s) = 1 = .
Ls2 + Rs + c Ls2 + Rs + 1c

Then from equation (9), I(s) = sQ(s) − q(0) implies

 
(Ls + R)q(0) + Li(0) + E(s)
I(s) = s − q(0)
Ls2 + Rs + 1c
Ls2 q(0) + Rsq(0) + Lsi(0) + sE(s) − Ls2 q(0) − Rsq(0) − 1c q(0)
=
Ls2 + Rs + 1c
Lsi(0) + sE(s) − 1c q(0)
= .
Ls2 + Rs + 1c

23. The equation is y 00 +16y = f (t); applying the Laplace transform we get L(y 00 )+16L(y) =
L(f (t)), so [s2 L(y) − sy(0) − y 0 (0)] + 16L(y) = L(f (t)). Now
Z ∞ Z π
−st 1 e−πs
L(f (t)) = f (t)e dt = e−st dt = − .
0 0 s s
Thus
1 e−πs
[s2 L(y) − 9s − 2] + 16L(y) = − ,
s s
which implies that
1 − e−πs
[s2 + 16]L(y) − 9s − 2 = ,
s
and we obtain
9s2 + 2s + 1 − e−πs
L(y) = .
s(s2 + 16)
5.2. PROPERTIES OF THE LAPLACE TRANSFORM 301

24. The equation is y 00 + y = f (t); applying the Laplace transform we get L(y 00 ) + L(y) =
L(f (t)), so [s2 L(y) − sy(0) − y 0 (0)] + L(y) = L(f (t)). Now
∞ 1
e−s e−s
Z Z
−st 1
L(f (t)) = f (t)e dt = te−st dt = − − 2 .
0 0 s2 s s

Thus
2 1 e−s e−s
s L(y) + L(y) = 2 − − 2 ,
s s s
which implies
1 − se−s − e−s
[s2 + 1]L(y) = ,
s2
and we obtain
1 − se−s − e−s
L(y) = .
s2 (s2 + 1)

25. The equation is y 00 + 4y = f (t); applying the Laplace transform we get L(y 00 ) + 4L(y) =
L(f (t)), so [s2 L(y) − sy(0) − y 0 (0)] + 4L(y) = L(f (t)). Now
∞ 1 ∞
1 − e−s
Z Z Z
−st −st
L(f (t)) = f (t)e dt = te dt + e−st dt = .
0 0 1 s2

Thus
1 − e−s
s2 L(y) + 4L(y) = ,
s2
which implies
1 − e−s
[s2 + 4]L(y) = ,
s2
and we obtain
1 − e−s
L(y) = .
s2 (s2 + 4)

26. The equation describing the amount of salt y(t) in the water at time t is given by

dy
= rate in − rate out.
dt
Here, the rate in = (1/2) lb/gal 8 gal/min = 4 gal/min. The rate out = (y(t)/640) lb/gal
8 gal/min = y/80 gal/min. Therefore, for the first 15 minutes, the equation describing the
amount of salt in the mixture is given by

dy y
+ =4 0 ≤ t ≤ 15.
dt 80
Then after 15 minutes, the rate of salt in becomes 0 gal/min. The rate of salt out remains
the same. Thus
dy y
+ =0 15 < t < ∞.
dt 80
302 CHAPTER 5. THE LAPLACE TRANSFORM

Combining these two equations, we have


dy y 4 0 ≤ t ≤ 15
+ = f (t) =
dt 80 0 15 < t < ∞.

Since the tank is assumed to initially contain fresh water, our initial condition is y(0) =
0. Applying the Laplace transform to the equation y 0 + y/80 = f (t), we have L(y 0 ) +
(1/80)L(y) = L(f (t)), so [sL(y) − y(0)] + (1/80)L(y) = L(f (t)). Now
∞ 15
1 − e−15s
Z Z
−st
L(f (t)) = f (t)e dt = 4e−st dt = 4 .
0 0 s

Therefore,
1 1 − e−15s
sL(y) + L(y) = 4 ,
80 s
which implies
1 1 − e−10s
[s + ]L(y) = 4 ,
80 s
and we obtain
1 − e−10s
L(y) = 4 .
s(s + 1/80)

27. The equation of motion is given by


00 0 F0 t 0≤t≤T
my + γy + ky = f (t) =
0 T < t < ∞.

Since the mass is initially in the equilibrium state, the initial conditions are y(0) = 0 and
y 0 (0) = 0. Applying the Laplace transform to the equation my 00 + γy 0 + ky = f (t), we have
mL(y 00 ) + γL(y 0 ) + kL(y) = L(f (t)), so m[s2 L(y) − sy(0) − y 0 (0)] + γ[sL(y) − y(0)] + kL(y) =
L(f (t)). Now
∞ T
1 − sT e−sT − e−sT
Z Z
−st
L(f (t)) = f (t)e dt = F0 te−st dt = F0 .
0 0 s2

Thus
1 − sT e−sT − e−sT
ms2 L(y) + γsL(y) + kL(y) = F0 ,
s2
which implies
1 − sT e−sT − e−sT
[ms2 + γs + k]L(y) = F0 ,
s2
and then
1 − sT e−sT − e−sT
L(y) = F0 .
s2 (ms2 + γs + k)
5.2. PROPERTIES OF THE LAPLACE TRANSFORM 303

28.(a) We compute:
∞ ∞
∞X
(−1)n t2n+1 −st
Z Z
−st
L(sin t) = (sin t)e dt = e dt
0 0 n=0
(2n + 1)!
∞ n ∞ ∞
(−1)n
Z  
X (−1) X (2n + 1)!
2n+1 −st
= t e dt =
n=0 0 (2n + 1)! n=0
(2n + 1)! s2n+2
∞ ∞  n
(−1)n
 
X 1 X −1 1 1 1
= 2n+2
= 2 2
= 2 2
= 2 ,
n=0
s s n=0
s s 1 + 1/s s + 1

where we used the fact that the sum of the geometric series ∞ n
P
n=0 r = 1/(1 − r) for |r| < 1.
∞
X (−1)n t2n+1
(b) Since the Taylor series for sin t = , we see that the Taylor series for f (t)
n=0
(2n + 1)!
is given by
∞
X (−1)n t2n
.
n=0
(2n + 1)!
Then
∞ ∞
∞X ∞ ∞
(−1)n t2n −st X (−1)n
Z Z Z
−st
L(f (t)) = f (t)e dt = e dt = t2n e−st dt
0 0 n=0
(2n + 1)! n=0
(2n + 1)! 0
∞ ∞
(−1)n (−1)n
 
X (2n)! X
= = = arctan(1/s),
n=0
(2n + 1)! s2n+1 n=0
(2n + 1)s2n+1
∞
X (−1)n t2n+1
where we have used the fact that the Taylor series for arctan t = .
n=0
(2n + 1)
(c) We compute:
∞ ∞
∞X
(−1)n t2n −st
Z Z
−st
L(J0 (t)) = J0 (t)e dt = e dt
0 0 n=0
22n (n!)2
∞ Z ∞ ∞
(−1)n (−1)n (2n)!
X X  
2n −st
= 2n 2
t e dt = 2n 2 2n+1
= (s2 + 1)−1/2 ,
n=0
2 (n!) 0 n=0
2 (n!) s

and
∞
√ Z ∞ √ −st Z
(−1)n tn −st ∞X
L(J0 ( t)) = J0 ( t)e dt = e dt
0 0 n=0
22n (n!)2
∞ Z ∞ ∞ ∞
(−1)n (−1)n (−1)n e−1/4s
  X
X
n −st
X n!
= 2n (n!)2
t e dt = 2n (n!)2 sn+1
= 2n (n!)sn+1
= .
n=0
2 0 n=0
2 n=0
2 s

29.(a) Taking the Laplace transform of the Airy equation, we have

L(y 00 ) − L(ty) = 0.
304 CHAPTER 5. THE LAPLACE TRANSFORM

Using the differentiation property of the transform, we get

d
L(y 00 ) + L(y) = 0.
ds
Therefore,
d
[s2 Y (s) − sy(0) − y 0 (0)] + Y (s) = 0.
ds
Using the initial conditions, we obtain
d
[s2 Y (s) − s] + Y (s) = 0,
ds
which we can rewrite as
Y 0 (s) + s2 Y (s) = s.

(b) Taking the Laplace transform of the Legendre equation, we have

L(y 00 ) − L(t2 y 00 ) − 2L(ty 0 ) + α(α + 1)L(y) = 0.

Using the differentiation property of the transform, we get

d2 d
L(y 00 ) − 2
L(y 00
) + 2 L(y 0 ) + α(α + 1)L(y) = 0.
ds ds
Therefore,
d2 2 d
[s2 Y (s)−sy(0)−y 0 (0)]− 2
[s Y (s)−sy(0)−y 0 (0)]++2 [sY (s)−y(0)]+α(α+1)Y (s) = 0.
ds ds
Using the initial conditions, we obtain
d2 2 d
s2 Y (s) − 1 − 2
[s Y (s) − 1] + 2 [sY (s)] + α(α + 1)Y (s) = 0.
ds ds
After carrying out the differentiation, the equation simplifies to

s2 Y 00 + 2sY 0 − [s2 + α(α + 1)]Y = −1.

5.3 The Inverse Laplace Transform

1. We need to find constants a, b such that

a(s − 4) + b(s + 3) = 3s − 40,

which is the same as

(a + b)s − 4a + 3b = 3s − 40.
Equating like coefficients, we need a + b = 3 and −4a + 3b = −40. Solving the first
equation, we have a = 3 − b. Plugging this equation into the second equation, we have
−4(3 − b) + 3b = −40. Therefore, b = −4 and a = 7.
5.3. THE INVERSE LAPLACE TRANSFORM 305

2. We need a1 , a2 to satisfy the following equation:

a1 (s + 2) + a2 = 3s + 4.

Equating like coefficients, we need a1 = 3 and 2a1 + a2 = 4. Therefore, a1 = 3 and a2 = −2.

3. We need a, b, c to satisfy the following equation,

a(s2 + 4) + (bs + 2c)(s + 4) = −3s2 + 32 − 14s.

We can rewrite this equation as

(a + b)s2 + (4b + 2c)s + 4a + 8c = −3s2 + 32 − 14s.

Equating like coefficients, we need a + b = −3, 4b + 2c = −14 and 4a + 8c = 32. Solving this
system of equations, we have a = 2, b = −5 and c = 3.
4. Dividing on the left side each term by the denominator, we get a1 = 3, a2 = 2 and a3 = 2.
5. We need a, b, c to satisfy the equation

a(s2 − 2s + 5) + (b(s − 1) + 2c)(s + 1) = 3s2 − 8s + 5.

This equation can be rewritten as

(a + b)s2 + (−2a + 2c)s + (5a − b + 2c) = 3s2 − 8s + 5.

Equating like coefficients, we have a + b = 3, −2a + 2c = −8 and 5a − b + 2c = 5. Solving

this system of equations, we have a = 2, b = 1 and c = −2.
6. We need a1 , a2 , b and c to satisfy the following equation:

a1 s(s + 3)(s + 1) + a2 (s + 3)(s + 1) + b(s + 1)s2 + c(s + 3)s2 = −2s3 − 8s2 + 8s + 6.

We can rewrite this equation as

(a1 + b + c)s3 + (4a1 + a2 + b + 3c)s2 + (3a1 + 4a2 )s + 3a2 = −2s3 − 8s2 + 8s + 6.

Equating like coefficients, we have 3a2 = 6 which implies a2 = 2. Then 3a1 + 4a2 = 8
which implies a1 = 0. Using this values for a1 and a2 and plugging back into the equations
a1 + b + c = −2 and 4a1 + a2 + b + 3c = −8, we conclude that c = −4 and b = 2. Therefore,
we have a1 = 0, a2 = 2, b = 2 and c = −4.
7. We need a1 , b1 , a2 , b2 , a3 , b3 to satisfy the equation

(a1 s + b1 )(s2 + 1)2 + (a2 s + b2 )(s2 + 1) + (a3 s + b3 ) = 8s3 − 15s − s5 .

We can rewrite this equation as

a1 s5 + b1 s4 + (2a1 + a2 )s3 + (2b1 + b2 )s2 + (a1 + a2 + a3 )s + (b1 + b2 + b3 ) = −s5 + 8s3 − 15s.

Equating like coefficients, we have a1 = −1 and b1 = 0. Then, 2a1 + a2 = 8 implies a2 = 10,

while 2b1 + b2 = 0 implies b2 = 0. Finally, a1 + a2 + a3 = −15 implies a3 = −24 and
306 CHAPTER 5. THE LAPLACE TRANSFORM

b1 + b2 + b3 = 0 implies b3 = 0. We conclude that a1 = −1, b1 = 0, a2 = 10, b2 = 0, a3 = −24

and b3 = 0.
8. We need a1 , b1 , a2 , b2 to satisfy the equation

(a1 (s + 1) + 2b1 )((s + 1)2 + 4) + a2 (s + 1) + 2b2 = s3 + 3s2 + 3s + 1.

We can rewrite this equation as

a1 s3 + (3a1 + 2b1 )s2 + (7a1 + 4b1 + a2 )s + (5a1 + 10b1 + a2 + 2b2 ) = s3 + 3s2 + 3s + 1.

Equating like coefficients, we have a1 = 1. Then 3a1 + 2b1 = 3 implies b1 = 0. Then

7a1 + 4b1 + a2 = 3 implies a2 = −4. Then 5a1 + 10b1 + a2 + 2b2 = 1 implies b2 = 0. To
summarize, we have a1 = 1, b1 = 0, a2 = −4 and b2 = 0.
9. Writing the function as
30 5
= 6 ,
s2 + 25 s2 + 5 2
we see that L−1 (Y (s)) = 6 sin 5t.
10. Writing the function as
4 2
3
=2 ,
(s − 3) (s − 3)3
we see that L−1 (Y (s)) = 2t2 e3t .
11. Using partial fractions, we write
 
2 2 1 1
= − .
s2 + 3s − 4 5 s−1 s+4
2
Thus we see that L−1 (Y (s)) = (et − e−4t ).
5
12. Using partial fractions, we write
3s 9 1 6 1
= + .
s2 −s−6 5s−3 5s+2
9 6
Thus we see that L−1 (Y (s)) = e3t + e−2t .
5 5
13. Completing the square in the denominator, we have
5s + 25 5(s + 5)
= .
s2 + 10s + 74 (s + 5)2 + 49

Thus we see that L−1 (Y (s)) = 5e−5t cos 7t.

14. We rewrite
6s − 3 s 3 2
2
=6 2 − 2 .
s −4 s −4 2s −4
3
Therefore, we see that L−1 (Y (s)) = 6 cosh 2t − sinh 2t.
2
5.3. THE INVERSE LAPLACE TRANSFORM 307

15. Completing the square in the denominator, we have

2s + 1 2s + 1 2(s − 1) 3
= = + .
s2 − 2s + 2 2
(s − 1) + 1 (s − 1) + 1 (s − 1)2 + 1
2

Therefore, we see that L−1 (Y (s)) = 2et cos t + 3et sin t.

16. Using partial fractions, we write

9s2 − 12s + 28 1 s 2
2
=7 +2 2 −6 2 .
s(s + 4) s s +4 s +4

Therefore, we see that L−1 (Y (s)) = 7 + 2 cos 2t − 6 sin 2t.

17. Completing the square in the denominator, we have

1 − 2s 5 − 2(s + 2)
= .
s2 + 4s + 5 (s + 2)2 + 1

Therefore, we see that L−1 (Y (s)) = 5e−2t sin t − 2e−2t cos t.

18. Completing the square in the denominator, we have

2s − 3 2(s + 1) − 5
= .
s2 + 2s + 10 (s + 1)2 + 9

5
Therefore, we see that L−1 (Y (s)) = 2e−t cos 3t − e−t sin 3t.
3
19. First we write
3s + 2 a b c
= + + .
(s − 2)(s + 2)(s + 1) s−2 s+2 s+1

Then a, b, c must satisfy the equation

a(s + 2)(s + 1) + b(s − 2)(s + 1) + c(s − 2)(s + 2) = 3s + 2.

We can rewrite this equation as

(a + b + c)s2 + (3a − b)s + (2a − 2b − 4c) = 3s + 2.

Equating like coefficients, we have a + b + c = 0, 3a − b = 3 and 2a − 2b − 4c = 2. Solving

this system of equations, we have a = 2/3, b = −1 and c = 1/3. Therefore, we can write

3s + 2 2/3 1 1/3
= − + ,
(s − 2)(s + 2)(s + 1) s−2 s+2 s+1

which implies
2 3 1
F (s) = − + .
s−2 s+2 s+1
308 CHAPTER 5. THE LAPLACE TRANSFORM

By the linearity of L−1 , we have

     
−1 −1 1 −1 1 −1 1
L (F (s)) = 2L − 3L +L = 2e2t − 3e−2t + e−t .
s−2 s+2 s+1

20. First we write

s3 − 2s2 + 8 a b c d
2
= + 2+ + .
s (s − 2)(s + 2) s s s−2 s+2
Then a, b, c, d must satisfy the equation

as(s − 2)(s + 2) + b(s − 2)(s + 2) + cs2 (s + 2) + ds2 (s − 2) = s3 − 2s2 + 8.

We can rewrite this equation as

(a + c + d)s3 + (b + 2c − 2d)s2 + (−4a)s − 4b = s3 − 2s2 + 8.

Equating like coefficients, we have a + c + d = 1, b + 2c − 2d = −2, −4a = 0 and −4b = 8.

Solving this system of equations, we have a = 0, b = −2, c = 1/2 and d = 1/2. Therefore,
we can write
s3 − 2s2 + 8 −2 1/2 1/2
2
= 2 + + ,
s (s − 2)(s + 2) s s−2 s+2
which implies
−4 1 1
F (s) = 2
+ + .
s s−2 s+2
By the linearity of L−1 , we have
     
−1 −1 1 −1 1 −1 1
L (F (s)) = 4L +L +L = 4t + e2t + e−2t .
s2 s−2 s+2

21. First we write

s2 − 7s + 28 as + b c
2
= 2 + .
(s − 8s + 25)(s + 3) s − 8s + 25 s + 3
Then a, b, c must satisfy the equation

(as + b)(s + 3) + c(s2 − 8s + 25) = s2 − 7s + 28.

We can rewrite this equation as

(a + c)s2 + (3a + b − 8c)s + (3b + 25c) = s2 − 7s + 28.

Equating like coefficients, we have a + c = 1, 3a + b − 8c = −7, and 3b + 25c = 28. Solving

this system of equations, we have a = 0, b = 1 and c = 1. Therefore, we can write
s2 − 7s + 28 1 1
2
= 2 + ,
(s − 8s + 25)(s + 3) s − 8s + 25 s + 3
which implies
12 12
F (s) = + .
s2 − 8s + 25 s + 3
5.3. THE INVERSE LAPLACE TRANSFORM 309

By the linearity of L−1 , we have

       
−1 −1 12 −1 12 −1 12 −1 12
L (F (s)) = L +L =L +L
s2 − 8s + 25 s+3 (s − 4)2 + 32 s+3
   
3 1
= 4L−1 2 2
+ 12L−1 = 4e4t sin 3t + 12e−3t .
(s − 4) + 3 s+3

22. First we write

s3 + 3s2 + 4s + 3 as + b cs + d
2 2
= 2 + .
(s + 1)(s + 4) s + 1 s2 + 4
Then a, b, c, d must satisfy the equation

(as + b)(s2 + 4) + (cs + d)(s2 + 1) = s3 + 3s2 + 4s + 3.

This equation can be rewritten as

(a + c)s3 + (b + d)s2 + (4a + c)s + (4b + d) = s3 + 3s2 + 4s + 3.

Equating like coefficients, we have a + c = 1, b + d = 3, 4a + c = 1 and 4b + d = 3. Solving

this system of equations, we have a = 0, b = 0, c = 1, and d = 3. Therefore,
     
s+3 s 2
F (s) = 2 2 =2 2 +3 2 .
s +4 s + 22 s + 22
By the linearity of L−1 , we have
   
−1 −1 s −1 2
L (F (s)) = 2L + 3L = 2 cos 2t + 3 sin 2t.
s2 + 2 2 s2 + 2 2

23. First we write

s3 − 2s2 − 6s − 6 as + b c d
2 2
= 2 + + 2.
(s + 2s + 2)s s + 2s + 2 s s
Then a, b, c, d must satisfy the equation

(as + b)s2 + cs(s2 + 2s + 2) + d(s2 + 2s + 2) = s3 − 2s2 − 6s − 6.

This equation can be rewritten as

(a + c)s3 + (b + 2c + d)s2 + (2c + 2d)s + 2d = s3 − 2s2 − 6s − 6.

Equating like coefficients, we have a + c = 1, b + 2c + d = −2, 2c + 2d = −6 and 2d = −6.

Solving this system of equations, we have a = 1, b = 1, c = 0, and d = −3. Therefore,
s+1 3 s+1 3
F (s) = − 2 = − 2.
s2 + 2s + 2 s 2
(s + 1) + 1 s
By the linearity of L−1 , we have
   
−1 −1 s+1 −1 1
L (F (s)) = L − 3L = e−t cos t − 3t.
(s + 1)2 + 1 s2
310 CHAPTER 5. THE LAPLACE TRANSFORM

24. First, we write

s2 + 3 as + b cs + d
2 2
= 2 + 2 .
(s + 2s + 2) s + 2s + 2 (s + 2s + 2)2

Then a, b, c, d must satisfy the equation

(as + b)(s2 + 2s + 3) + cs + d = s2 + 3.

We can rewrite this equation as

as3 + (2a + b)s2 + (2a + 2b + c)s + (2b + d) = s2 + 3.

Equating like coefficients, we have a = 0, 2a + b = 1, 2a + 2b + c = 0 and 2b + d = 3. Solving

this system of equations, we have a = 0, b = 1, c = −2 and d = 1. Therefore,

1 −2s + 1
F (s) = + 2 .
s2 + 2s + 2 (s + 2s + 2)2

We can check that

 
d As + B A 4A − 2B + (2A − 2B)s
2
=− 2 + .
ds s + 2s + 2 s + 2s + 2 (s2 + 2s + 2)2

Solving the equations 4A − 2B = 1, 2A − 2B = −2 we obtain A = 3/2, B = 5/2. Thus (also

completing the square in the denominator) we can write
 
d (3/2)s + (5/2) 5/2
F (s) = 2
+ .
ds (s + 1) + 1 (s + 1)2 + 1

Now
(3/2)s + (5/2) 3 s+1 1
2
= + ,
(s + 1) + 1 2 (s + 1) + 1 (s + 1)2 + 1
2

thus we obtain that

3 5
L−1 (F (s)) = − te−t cos t − te−t sin t + e−t sin t.
2 2

25. Using a computer algebra system, we obtain that

   
−1 7 −t 8 4
L (Y (s)) = 1 − t e cos 3t − + t e−t sin 3t.
3 9 3

26. Using a computer algebra system, we obtain that

 
−1 591 367 51 2 4 3 1 4 591 3t 31 3t
L (Y (s)) = − + t+ t + t + t + e cos t + e sin t.
6250 1250 125 25 40 6250 3125
5.4. SOLVING DIFFERENTIAL EQUATIONS WITH LAPLACE TRANSFORMS 311

27. Using a computer algebra system, we obtain that

   
−1 1 13 1 2 t 261 3 1 2 t
L (Y (s)) = + t − t e cos t + − + t + t e sin t
250 25 40 500 200 5
 
1 1
− + t e−2t .
250 100

28. Using a computer algebra system, we obtain that

   
−1 59 21 3 2 3t 59 1
L (Y (s)) = − − t− t e + + t e−t cos 2t
4000 400 160 4000 80
 
3 7
− + t e−t sin 2t.
1000 320

5.4 Solving Differential Equations with Laplace Trans-

forms
1. Applying the Laplace transform to the differential equation, we get L(y 00 ) − 4L(y 0 ) −
12L(y) = 0, which is [s2 L(y) − sy(0) − y 0 (0)] − 4[sL(y) − y(0)] − 12L(y) = 0. Using the
initial conditions, this gives [s2 L(y) − 8s] − 4[sL(y) − 8] − 12L(y) = 0, which turns into
[s2 − 4s − 12]L(y) = 8s − 32, thus (using partial fractions)
8s − 32 2 6
L(y) = = + ,
s2 − 4s − 12 s−6 s+2
which implies that
y(t) = 2e6t + 6e−2t .

2. Applying the Laplace transform to the differential equation, we get L(y 00 ) + 3L(y 0 ) +
2L(y) = L(t), which is [s2 L(y) − sy(0) − y 0 (0)] + 3[sL(y) − y(0)] + 2L(y) = 1/s2 . Using the
initial conditions, this gives [s2 L(y) − s] + 3[sL(y) − 1] + 2L(y) = 1/s2 , which turns into
[s2 + 3s + 2]L(y) = s + 3 + 1/s2 , and then
s+3 1 s3 + 3s2 + 1
L(y) = + = .
s2 + 3s + 2 s2 (s2 + 3s + 2) s2 (s + 2)(s + 1)
Using partial fractions, we write
s3 + 3s2 + 1 a b c d
L(y) = 2
= + 2+ + .
s (s + 2)(s + 1) s s s+2 s+1
Then a, b, c, d must satisfy the equation

as(s + 2)(s + 1) + b(s + 2)(s + 1) + cs2 (s + 1) + ds2 (s + 2) = s3 + 3s2 + 1.

This equation can be rewritten as

(a + c + d)s3 + (3a + b + c + 2d)s2 + (2a + 3b)s + 2b = s3 + 3s2 + 1.

312 CHAPTER 5. THE LAPLACE TRANSFORM

Equating like coefficients, we have a + c + d = 1, 3a + b + c + 2d = 3, 2a + 3b = 0 and

2b = 1. Solving this system of equations, we have a = −3/4, b = 1/2, c = −5/4 and d = 3.
Therefore,
−3/4 1/2 −5/4 3
L(y) = + 2 + + ,
s s s+2 s+1
which implies that
       
3 −1 1 1 −1 1 5 −1 1 −1 1
y = − L + L − L + 3L
4 s 2 s2 4 s+2 s+1
3 1 5
= − + t − e−2t + 3e−t .
4 2 4

3. Applying the Laplace transform to the differential equation, we get L(y 00 ) − 8L(y 0 ) +
25L(y) = 0, which is [s2 L(y) − sy(0) − y 0 (0)] − 8[sL(y) − y(0)] + 25L(y) = 0. Using
the initial conditions, this gives [s2 L(y) − 3] − 8sL(y) + 25L(y) = 0, which turns into
[s2 − 8s + 25]L(y) = 3, and then
3
L(y) = .
s2 − 8s + 25
Completing the square in the denominator, we have
3
L(y) = ,
(s − 4)2 + 9
which implies that
y(t) = e4t sin 3t.

4. Applying the Laplace transform to the differential equation, we get L(y 00 ) − 4L(y 0 ) +
4L(y) = 0, which is [s2 L(y) − sy(0) − y 0 (0)] − 4[sL(y) − y(0)] + 4L(y) = 0. Using the
initial conditions, this gives [s2 L(y) − s − 1] − 4[sL(y) − 1] + 4L(y) = 0, which turns into
[s2 − 4s + 4]L(y) = s − 3, and then
s−3
L(y) = .
s2 − 4s + 4
Therefore, we have
s−3 s−2 1 1 1
L(y) = = − = − ,
(s − 2)2 (s − 2)2 (s − 2)2 s − 2 (s − 2)2
which implies that
y(t) = e2t − te2t .

5. Applying the Laplace transform to the differential equation, we get L(y 00 ) − 2L(y 0 ) +
4L(y) = 0, which is [s2 L(y) − sy(0) − y 0 (0)] − 2[sL(y) − y(0)] + 4L(y) = 0. Using the
initial conditions, this gives [s2 L(y) − 2s] − 2[sL(y) − 2] + 4L(y) = 0, which turns into
[s2 − 2s + 4]L(y) = 2s − 4, and then
2s − 4
L(y) = .
s2 − 2s + 4
5.4. SOLVING DIFFERENTIAL EQUATIONS WITH LAPLACE TRANSFORMS 313

Completing the square in the denominator, we have

2s − 4 2(s − 1) 2
L(y) = = − ,
(s − 1)2 + 3 (s − 1)2 + 3 (s − 1)2 + 3
which implies that √
√ 2 3 t √
y(t) = 2et cos( 3 t) − e sin( 3 t).
3
6. Applying the Laplace transform to the differential equation, we get L(y 00 ) + 4L(y 0 ) +
29L(y) = L(e−2t sin 5t), which is
5
[s2 L(y) − sy(0) − y 0 (0)] + 4[sL(y) − y(0)] + 29L(y) = .
(s + 2)2 + 25
Using the initial conditions, this gives
5
[s2 L(y) − 5s + 2] + 4[sL(y) − 5] + 19L(y) = ,
(s + 2)2 + 25
which turns into
5
[s2 + 4s + 19]L(y) − 5s + 2 − 20 = ,
(s + 2)2 + 25
and then
5s + 18 5 5s + 18 5
L(y) = + = + .
s2 + 4s + 29 (s2 + 4s + 29)[(s + 2)2 + 25] s2 + 4s + 29 (s2 + 4s + 29)2
We can check that
 
d As + B A 58A − 4B + (4A − 2B)s
2
=− 2 + .
ds s + 4s + 29 s + 4s + 29 (s2 + 4s + 29)2

Solving the equations 58A − 4B = 5, 4A − 2B = 0 we obtain A = 1/10, B = 1/5. Thus

(also completing the square in the denominator) we can write
 
d s/10 + 1/5 1/10 5s + 18
L(y) = 2
+ +
ds (s + 2) + 25 (s + 2) + 25 (s + 2)2 + 25
2
 
d s/10 + 1/5 5(s + 2) 81/10
= 2
+ + .
ds (s + 2) + 25 (s + 2) + 25 (s + 2)2 + 25
2

Therefore, we conclude that

1 −2t 81
y= te cos 5t + 5e−2t cos 5t + e−2t sin 5t.
10 50

7. Applying the Laplace transform to the differential equation, we get L(y 00 ) + ω 2 L(y) =
L(cos 2t), which is
s
[s2 L(y) − sy(0) − y 0 (0)] + ω 2 L(y) = 2 .
s +4
314 CHAPTER 5. THE LAPLACE TRANSFORM

Using the initial conditions, this gives

s
[s2 L(y) − s] + ω 2 L(y) = ,
s2 +4
which turns into
s
[s2 + ω 2 ]L(y) = s + ,
s2 +4
and then
s s
L(y) = + 2 .
s2 +ω 2 (s + ω )(s2 + 4)
2

Using partial fractions in the second term, we rewrite L(y) as

 
s 1 s s
L(y) = 2 + − ,
s + ω 2 4 − ω 2 s2 + ω 2 s2 + 4

which implies
1 1
y = cos ωt + 2
[cos ωt − cos 2t] = [(5 − ω 2 ) cos ωt − cos 2t].
4−ω 4 − ω2

8. Applying the Laplace transform to the differential equation, we get L(y 00 ) − 2L(y 0 ) +
2L(y) = L(cos t), which is
s
[s2 L(y) − sy(0) − y 0 (0)] − 2[sL(y) − y(0)] + 2L(y) = .
s2 +1
Using the initial conditions, this gives
s
[s2 L(y) − s] − 2[sL(y) − 1] + 2L(y) = ,
s2 +1
which turns into
s
[s2 − 2s + 2]L(y) = s + 2 + ,
s2 +1
and then
s+2 s
L(y) = + .
s2 − 2s + 2 (s2 − 2s + 2)(s2 + 1)
Using partial fractions in the second term, we get
 
s 1 s−2 s−4
= − .
(s2 − 2s + 2)(s2 + 1) 5 s2 + 1 s2 − 2s + 2

Thus
1 s 2 1 2 2s − 3
L(y) = 2
− 2 + 2 .
5 s + 1 5 s + 1 5 s − 2s + 2
Completing the square in the denominator for the last term, we have

2s − 3 2(s − 1) − 1
= .
s2 − 2s + 2 (s − 1)2 + 1
5.4. SOLVING DIFFERENTIAL EQUATIONS WITH LAPLACE TRANSFORMS 315

Therefore,
1 s 2 1 2 2(s − 1) − 1
L(y) = − 2 + ,
5 s + 1 5 s + 1 5 (s − 1)2 + 1
2

which implies
1 2 4 2
y= cos t − sin t + et cos t − et sin t.
5 5 5 5
9. Applying the Laplace transform to the differential equation, we get L(y 00 ) − 2L(y 0 ) +
2L(y) = L(e−t ), which is
1
[s2 L(y) − sy(0) − y 0 (0)] − 2[sL(y) − y(0)] + 2L(y) = .
s+1
Using the initial conditions, this gives
1
[s2 L(y) − 1] − 2[sL(y)] + 2L(y) = ,
s+1
which turns into
1
[s2 − 2s + 2]L(y) = 1 + ,
s+1
and then
1 1
L(y) = + 2 .
s2 − 2s + 2 (s − 2s + 2)(s + 1)
Using partial fractions in the second term, we get
1 1 1 1 3−s
= + 2 .
(s2 − 2s + 2)(s + 1) 5 s + 1 5 s − 2s + 2

Thus
1 1 1 8−s
L(y) = + 2 .
5 s + 1 5 s − 2s + 2
Completing the square in the denominator for the last term, we have

8−s (s − 1) − 7
=− .
s2 − 2s + 2 (s − 1)2 + 1

Therefore,
1 1 1 (s − 1) − 7
L(y) = − ,
5 s + 1 5 (s − 1)2 + 1
which implies
1 1 7
y = e−t − et cos t + et sin t.
5 5 5
10. Applying the Laplace transform to the differential equation, we get L(y 00 ) + 2L(y 0 ) +
L(y) = 18L(e−t ), which is
18
[s2 L(y) − sy(0) − y 0 (0)] + 2[sL(y) − y(0)] + L(y) = .
s+1
316 CHAPTER 5. THE LAPLACE TRANSFORM

Using the initial conditions, this gives

18
[s2 L(y) − 7s + 2] + 2[sL(y) − 7] + L(y) = ,
s+1
which turns into
18
[s2 + 2s + 1]L(y) = 7s + 12 + ,
s+1
and then
7s + 12 18 7(s + 1) + 5 18
L(y) = + 2 = +
s2
+ 2s + 1 (s + 2s + 1)(s + 1) (s + 1) 2 (s + 1)3
7 5 18
= + 2
+ .
s + 1 (s + 1) (s + 1)3
Therefore,
y = 7e−t + 5te−t + 9t2 e−t .

11. Applying the Laplace transform to the differential equation, we get L(y (4) ) − 4L(y 000 ) +
6L(y 00 ) − 4L(y 0 ) + L(y) = 0, which is [s4 L(y) − s3 y(0) − s2 y 0 (0) − sy 00 (0) − y 000 (0)] − 4[s3 L(y) −
s2 y(0)−sy 0 (0)−y 00 (0)]+6[s2 L(y)−sy(0)−y 0 (0)]−4[sL(y)−y(0)]+L(y) = 0. Using the initial
conditions, this gives [s4 L(y) − s2 − 1] − 4[s3 L(y) − s] + 6[s2 L(y) − 1] − 4[sL(y)] + L(y) = 0,
which turns into [s4 − 4s3 + 6s2 − 4s + 1]L(y) − s2 − 1 + 4s − 6 = 0, and then

s2 − 4s + 7 s2 − 4s + 7
L(y) = = .
s4 − 4s3 + 6s2 − 4s + 1 (s − 1)4
Using partial fractions, we write

s2 − 4s + 7 a b c d
L(y) = = + + + .
(s − 1)4 s − 1 (s − 1)2 (s − 1)3 (s − 1)4
Then a, b, c, d must satisfy the equation

a(s − 1)3 + b(s − 1)2 + c(s − 1) + d = s2 − 4s + 7.

This equation can be rewritten as

as3 + (−3a + b)s2 + (3a − 2b + c)s + (−a + b − c + d) = s2 − 4s + 7.

Equating like coefficients, we have a = 0, −3a+b = 1, 3a−2b+c = −4 and −a+b−c+d = 7.

Solving this system of equations, we have a = 0, b = 1, c = −2 and d = 4. Thus
1 2 4
L(y) = 2
− 3
+ ,
(s − 1) (s − 1) (s − 1)4
which implies that
     
−1 1 −1 2 2 −1 6 2
y=L 2
−L 3
+ L 4
= tet − t2 et + t3 et .
(s − 1) (s − 1) 3 (s − 1) 3
5.4. SOLVING DIFFERENTIAL EQUATIONS WITH LAPLACE TRANSFORMS 317

12. Applying the Laplace transform to the differential equation, we get L(y (4) ) − L(y) = 0,
which is [s4 L(y) − s3 y(0) − s2 y 0 (0) − sy 00 (0) − y 000 (0)] − L(y) = 0. Using the initial conditions,
this gives [s4 L(y) − s3 − s] − L(y) = 0, which turns into [s4 − 1]L(y) = s3 + s, and then

s3 + s s3 + s s
L(y) = 4
= 2 2
= 2 .
s −1 (s − 1)(s + 1) s −1
Using partial fractions, we write
s a b
L(y) = = + .
s2 − 1 s−1 s+1
Then a, b must satisfy the equation a(s + 1) + b(s − 1) = s. This equation can be rewritten
as (a + b)s + (a − b) = s. Equating like coefficients, we have a + b = 1, a − b = 0. Solving
this system of equations, we have a = 1/2 and b = 1/2. Thus

1/2 1/2
L(y) = + ,
s−1 s+1
which implies    
1 1 1 1 1 1
y = L−1 + L−1 = et + e−t .
2 s−1 2 s+1 2 2

13. Applying the Laplace transform to the differential equation, we get L(y (4) ) − 9L(y) = 0,
which is [s4 L(y) − s3 y(0) − s2 y 0 (0) − sy 00 (0) − y 000 (0)] − 9L(y) = 0. Using the initial conditions,
this gives [s4 L(y) − s3 + 3s] − 9L(y) = 0, which turns into (s4 − 9)L(y) = s3 − 3s, and then

s3 − 3s s
L(y) = 4
= 2 .
s −9 s +3
Thus
√
 
−1 s
y=L 2
= cos( 3 t).
s +3
14. The system can be written as

y10 = −5y1 + y2
y20 = −9y1 + 5y2

with the initial conditions y1 (0) = 1 and y2 (0) = 0. Applying the Laplace transform to each
equation, we get

L(y10 ) = −5L(y1 ) + L(y2 )

L(y20 ) = −9L(y1 ) + 5L(y2 ),

which implies

sL(y1 ) − y1 (0) = −5L(y1 ) + L(y2 )

sL(y2 ) − y2 (0) = −9L(y1 ) + 5L(y2 ).
318 CHAPTER 5. THE LAPLACE TRANSFORM

Letting Y1 = L(y1 ), Y2 = L(y2 ) and plugging in the initial conditions, we obtain

sY1 − 1 = −5Y1 + Y2
sY2 − 0 = −9Y1 + 5Y2 .

These equations can be written in matrix form as

    
s + 5 −1 Y1 1
= .
9 s−5 Y2 0

The solution of this system is given by

      
Y1 1 s−5 1 1 1 s−5
= = 2 .
Y2 (s + 5)(s − 5) + 9 −9 s + 5 0 s − 16 −9

s−5
Thus Y1 (s) = , and then
s2 − 16
   
−1 s−5 −1 9/8 1/8 9 1
y1 (t) = L 2
=L − = e−4t − e4t ,
s − 16 s+4 s−4 8 8

9
and Y2 (s) = − , thus
s2− 16
   
−1 9 −1 9/8 9/8 9 9
y2 (t) = −L 2
= −L − = e−4t − e4t .
s − 16 s+4 s−4 8 8

15. The system can be written as

y10 = 5y1 − 2y2

y20 = 6y1 − 2y2

with the initial conditions y1 (0) = 1 and y2 (0) = 0. Applying the Laplace transform to each
equation, we have

L(y10 ) = 5L(y1 ) − 2L(y2 )

L(y20 ) = 6L(y1 ) − 2L(y2 ),

which implies

sL(y1 ) − y1 (0) = 5L(y1 ) − 2L(y2 )

sL(y2 ) − y2 (0) = 6L(y1 ) − 2L(y2 ).

Letting Y1 = L(y1 ), Y2 = L(y2 ) and plugging in the initial conditions, we have

sY1 − 1 = 5Y1 − 2Y2

sY2 − 0 = 6Y1 − 2Y2 .
5.4. SOLVING DIFFERENTIAL EQUATIONS WITH LAPLACE TRANSFORMS 319

These equations can be written in matrix form as

    
s−5 2 Y1 1
= .
−6 s + 2 Y2 0

The solution of this system is given by

      
Y1 1 s + 2 −2 1 1 s+2
= = 2 .
Y2 (s − 5)(s + 2) + 12 6 s−5 0 s − 3s + 2 6

We first consider Y1 (s). Using partial fractions, we write

s+2 s+2 a b
Y1 (s) = = = + .
s2 − 3s + 2 (s − 2)(s − 1) s−2 s−1

Then a, b must satisfy the equation a(s − 1) + b(s − 2) = s + 2. We can simplify this equation,
rewriting it as (a + b)s + (−a − 2b) = s + 2. Equating like coefficients, we have a + b = 1
and −a − 2b = 2. The solution of this system is a = 4 and b = −3. Thus
4 3
Y1 (s) = − ,
s−2 s−1
and then    
−1 1 −1 1
y1 (t) = 4L − 3L = 4e2t − 3et .
s−2 s−1
Next, we consider Y2 (s). Using partial fractions, we write
6 6 6 6
Y2 (s) = = = − ,
s2 − 3s + 2 (s − 2)(s − 1) s−2 s−1
and then    
−1 1 −1 1
y2 (t) = 6L − 6L = 6e2t − 6et .
s−2 s−1
16. The system can be written as

y10 = 4y1 − 4y2

y20 = 5y1 − 4y2

with the initial conditions y1 (0) = 1 and y2 (0) = 0. Applying the Laplace transform to each
equation, we have

L(y10 ) = 4L(y1 ) − 4L(y2 )

L(y20 ) = 5L(y1 ) − 4L(y2 ),

which implies

sL(y1 ) − y1 (0) = 4L(y1 ) − 4L(y2 )

sL(y2 ) − y2 (0) = 5L(y1 ) − 4L(y2 ).
320 CHAPTER 5. THE LAPLACE TRANSFORM

Letting Y1 = L(y1 ), Y2 = L(y2 ) and plugging in the initial conditions, we have

sY1 − 1 = 4Y1 − 4Y2

sY2 − 0 = 5Y1 − 4Y2 .

These equations can be written in matrix form as

    
s−4 4 Y1 1
= .
−5 s + 4 Y2 0
The solution of this system is given by
      
Y1 1 s + 4 −4 1 1 s+4
= = 2 .
Y2 (s − 4)(s + 4) + 20 5 s−4 0 s +4 5
s+4
Now Y1 (s) = , thus
s2 + 4
   
−1 s −1 2
y1 (t) = L 2
+ 2L 2
= cos 2t + 2 sin 2t.
s +4 s +4
5
Next, Y2 (s) = , thus
s2 +4
 
5 2 5
y2 (t) = L−1 2
= sin 2t.
2 s +4 2

17. The system can be written as

y10 = 6y2 , y20 = −6y1

with the initial conditions y1 (0) = 5 and y2 (0) = 4. Applying the Laplace transform to each
equation, we have
L(y10 ) = 6L(y2 ), L(y20 ) = −6L(y1 ),
which implies

sL(y1 ) − y1 (0) = 6L(y2 ), sL(y2 ) − y2 (0) = −6L(y1 ).

Letting Y1 = L(y1 ), Y2 = L(y2 ) and plugging in the initial conditions, we have

sY1 − 5 = 6Y2 , sY2 − 4 = −6Y1 .

These equations can be written in matrix form as

    
s −6 Y1 5
= .
6 s Y2 4
The solution of this system is given by
      
Y1 1 s 6 5 1 5s + 24
= 2 = 2 .
Y2 s + 36 −6 s 4 s + 36 4s − 30
5.4. SOLVING DIFFERENTIAL EQUATIONS WITH LAPLACE TRANSFORMS 321

5s + 24
Now Y1 (s) = , thus
s2 + 36
   
−1 s −1 6
y1 (t) = 5L 2
+ 4L 2
= 5 cos 6t + 4 sin 6t.
s + 36 s + 36
4s − 30
Next, Y2 (s) = , thus
s2 + 36
   
−1 s −1 6
y2 (t) = 4L 2
− 5L 2
= 4 cos 6t − 5 sin 6t.
s + 36 s + 26

18. The system can be written as

y10 = −4y1 − y2
y20 = y1 − 2y2

with the initial conditions y1 (0) = 1 and y2 (0) = 0. Applying the Laplace transform to each
equation, we have

L(y10 ) = −4L(y1 ) − L(y2 )

L(y20 ) = L(y1 ) − 2L(y2 ),

which implies

sL(y1 ) − y1 (0) = −4L(y1 ) − L(y2 )

sL(y2 ) − y2 (0) = L(y1 ) − 2L(y2 ).

Letting Y1 = L(y1 ), Y2 = L(y2 ) and plugging in the initial conditions, we have

sY1 − 1 = −4Y1 − Y2
sY2 − 0 = Y1 − 2Y2 .

These equations can be written in matrix form as

    
s+4 1 Y1 1
= .
−1 s + 2 Y2 0

The solution of this system is given by

      
Y1 1 s + 2 −1 1 1 s+2
= = 2 .
Y2 (s + 4)(s + 2) + 1 1 s+4 0 s + 6s + 9 1

s+2 s+3−1 1 1
Now Y1 (s) = 2
= 2
= − , thus
(s + 3) (s + 3) s + 3 (s + 3)2
   
−1 1 −1 1
y1 (t) = L −L = e−3t − te−3t .
s+3 (s + 3)2
322 CHAPTER 5. THE LAPLACE TRANSFORM

1
Next, Y2 (s) = , thus
(s + 3)2
 
−1 1
y2 (t) = L = te−3t .
(s + 3)2

19. The system can be written as

y10 = 2y1 − 64y2

y20 = y1 − 14y2

with the initial conditions y1 (0) = 0 and y2 (0) = 1. Applying the Laplace transform to each
equation, we have

L(y10 ) = 2L(y1 ) − 64L(y2 )

L(y20 ) = L(y1 ) − 14L(y2 ),

which implies

sL(y1 ) − y1 (0) = 2L(y1 ) − 64L(y2 )

sL(y2 ) − y2 (0) = L(y1 ) − 14L(y2 ).

Letting Y1 = L(y1 ), Y2 = L(y2 ) and plugging in the initial conditions, we have

sY1 − 0 = 2Y1 − 64Y2

sY2 − 1 = Y1 − 14Y2 .

These equations can be written in matrix form as

    
s−2 64 Y1 0
= .
−1 s + 14 Y2 1

The solution of this system is given by

      
Y1 1 s + 14 −64 0 1 −64
= = 2 .
Y2 (s − 2)(s + 14) + 64 1 s−2 1 s + 12s + 36 s − 2
64
Now Y1 (s) = − , thus
(s + 6)2
 
−1 1
y1 (t) = −64L = −64te−6t .
(s + 6)2
s−2 s+6−8 1 8
Next, Y2 (s) = 2
= 2
= − , thus
(s + 6) (s + 6) s + 6 (s + 6)2
   
−1 1 −1 1
y2 (t) = L − 8L = e−6t − 8te−6t .
s+6 (s + 6)2
5.4. SOLVING DIFFERENTIAL EQUATIONS WITH LAPLACE TRANSFORMS 323

20. Applying the Laplace transform, we have

L(y10 ) = −4L(y1 ) − L(y2 ) + L(2et )

L(y20 ) = L(y1 ) − 2L(y2 ) + L(sin 2t).

Letting Y1 = L(y1 ) and Y2 = L(y2 ) and using the property of the transform of a derivative,
we have
2
sY1 − y1 (0) = −4Y1 − Y2 +
s−1
2
sY2 − y2 (0) = Y1 − 2Y2 + 2 .
s +4
Plugging in the initial conditions, we obtain
2
sY1 − 1 = −4Y1 − Y2 +
s−1
2
sY2 − 2 = Y1 − 2Y2 + 2 .
s +4
This system can be written in matrix form as
    s+1 
s+4 1 Y1 s−1
= 2s2 +10 .
−1 s + 2 Y2 s2 +4

The solution of this system is given by

!
  s4 +s3 +8s2 +2s+18
Y1 1 (s−1)(s2 +4)
= 2s4 +7s3 +3s2 +34s−36 .
Y2 (s + 3)2 (s−1)(s2 +4)

Taking the inverse Laplace transform, we obtain

 
69t 749 3 12 5
y1 (t) = − + e−3t + et + cos 2t − sin 2t
26 1352 8 169 169
and  
69t 2839 1 38 44
y2 (t) = + e−3t + et − cos 2t + sin 2t.
26 1352 8 169 169
21. Applying the Laplace transform, we have

L(y10 ) = 5L(y1 ) − L(y2 ) + L(e−t )

L(y20 ) = L(y1 ) + 3L(y2 ) + L(2et )

Letting Y1 = L(y1 ) and Y2 = L(y2 ) and using the property of the transform of a derivative,
we have
1
sY1 − y1 (0) = 5Y1 − Y2 +
s+1
2
sY2 − y2 (0) = Y1 + 3Y2 + .
s−1
324 CHAPTER 5. THE LAPLACE TRANSFORM

Plugging in the initial conditions, we obtain

1
sY1 + 3 = 5Y1 − Y2 +
s+1
2
sY2 − 2 = Y1 + 3Y2 + .
s−1
This system can be written in matrix form as
−3s−2
    
s−5 1 Y1 s+1
= 2s .
−1 s − 3 Y2 s−1

The solution of this system is given by

!
−3s3 +8s2 −3s−6
 
Y1 1 s2 −1
= 2s3 −11s2 −9s+2 .
Y2 (s − 4)2 s2 −1

Taking the inverse Laplace transform, we obtain

 
82t 589 4t 4 2
y1 (t) = − − e − e−t − et
15 225 25 9
and  
82t 641 4t 1 8
y2 (t) = − + e + e−t − et .
15 225 25 9
22. Applying the Laplace transform, we have

L(y10 ) = −L(y1 ) − 5L(y2 ) + L(3)

L(y20 ) = L(y1 ) + 3L(y2 ) + L(5 cos t).

Letting Y1 = L(y1 ) and Y2 = L(y2 ) and using the property of the transform of a derivative,
we get
3
sY1 − y1 (0) = −Y1 − 5Y2 +
s
5s
sY2 − y2 (0) = Y1 + 3Y2 + 2 .
s +1
Plugging in the initial conditions, we obtain
3
sY1 − 1 = −Y1 − 5Y2 +
s
5s
sY2 + 1 = Y1 + 3Y2 + 2 .
s +1
This system can be written in matrix form as
    s+3 
s+1 5 Y1 s
= 2
−s +5s−1 .
−1 s − 3 Y2 2
s +1
5.4. SOLVING DIFFERENTIAL EQUATIONS WITH LAPLACE TRANSFORMS 325

The solution of this system is given by

!
  s4 +5s3 −33s2 +5s−9
Y1 1 s(s2 +1)
= 2 −s4 +5s3 +7s2 +3 .
Y2 s − 2s + 2 s(s2 +1)

Taking the inverse Laplace transform, we obtain

9 21 27
y1 (t) = − + et cos t − et sin t − 5 cos t + 10 sin t
2 2 2
and
3 3 t 15
y2 (t) = − e cos t + et sin t − cos t − 3 sin t.
2 2 2
23. Applying the Laplace transform, we have

L(y10 ) = −2L(y1 ) + L(y2 )

L(y20 ) = L(y1 ) − 2L(y2 ) + L(sin t).

Letting Y1 = L(y1 ) and Y2 = L(y2 ) and using the property of the transform of a derivative,
we get

sY1 − y1 (0) = −2Y1 + Y2

1
sY2 − y2 (0) = Y1 − 2Y2 + .
s2 +1
Plugging in the initial conditions, we obtain

sY1 = −2Y1 + Y2
sY2 = Y1 − 2Y2 + s21+1 .

This system can be written in matrix form as

    
s + 2 −1 Y1 0
= 1 .
−1 s + 2 Y2 s2 +1

The solution of this system is given by

1
   
Y1 1 s2 +1
= 2 s+2 .
Y2 s + 4s + 3 s2 +1

Taking the inverse Laplace transform, we get

1 1 1 1
y1 (t) = e−t − e−3t − cos t + sin t
4 20 5 10
and
1 −3t 3 2 1
y2 (t) = e − cos t + sin t + e−t .
20 10 5 4
326 CHAPTER 5. THE LAPLACE TRANSFORM

24. Applying the Laplace transform, we have

L(y10 ) = L(y2 ) − L(y3 )

L(y20 ) = L(y1 ) + L(y3 ) − L(e−t )
L(y30 ) = L(y1 ) + L(y2 ) + L(et ).

Letting Y1 = L(y1 ), Y2 = L(y2 ), Y3 = L(y3 ) and using the property of the transform of a
derivative, we obtain

sY1 − y1 (0) = Y2 − Y3
1
sY2 − y2 (0) = Y1 + Y3 −
s+1
1
sY3 − y3 (0) = Y1 + Y2 + .
s−1
Plugging in the initial conditions, we get

sY1 − 1 = Y2 − Y3
1
sY2 − 2 = Y1 + Y3 −
s+1
1
sY3 − 3 = Y1 + Y2 + .
s−1
This system can be written in matrix form as
    
s −1 1 Y1 1
 −1 s −1   Y2  =  2s+1 .
s+1
3s−2
−1 −1 s Y3 s−1

The solution of this system is given by

  1 2s+1 3s−2

+ (s+1) 2 s − (s+1)s(s−1)

Y1 s
2 +1)(2s+1)
 Y2  =  1
+ (s 3s−2
+ (s+1)s(s−1) .

 (s−1)s s(s2 −1)(s+1)
Y3 1
+ 2s+1
+ 3s−2
(s−1)s (s−1)s(s+1) s(s−1)

Taking the inverse Laplace transform, we obtain

3 1
y1 (t) = te−t + e−t − et ,
2 2
and
y2 (t) = −te−t − e−t + 3et ,
and
1 7
y3 (t) = − e−t + et .
2 2
25. The system cannot have a solution, because the initial conditions x0 (0) = 1, y 0 (0) = 2
do not satisfy the second equation at t = 0.
5.4. SOLVING DIFFERENTIAL EQUATIONS WITH LAPLACE TRANSFORMS 327

26. Applying the Laplace transform to each equation, we have

L(m01 ) = −k1 L(m1 )

L(m02 ) = k1 L(m1 ) − k2 L(m2 )
L(m03 ) = k2 L(m2 ).

Letting M1 = L(m1 ), M2 = L(m2 ), M3 = L(m3 ) and using the property involving the
Laplace transform of a derivative, we obtain

sM1 − m1 (0) = −k1 M1

sM2 − m2 (0) = k1 M1 − k2 M2
sM3 − m3 (0) = k2 M2 .

Plugging in the initial conditions, we get

sM1 − m0 = −k1 M1
sM2 = k1 M1 − k2 M2
sM3 = k2 M2 .

The first equation implies that (s + k1 )M1 = m0 , thus

m0
M1 = ,
s + k1

and then  
−1 1
m1 (t) = m0 L = m0 e−k1 t .
s + k1
The second equation gives
k1 m0
(s + k2 )M2 = ,
s + k1
thus
k1 m0
M2 = .
(s + k2 )(s + k1 )
Using partial fractions, we have

a b
M2 = + .
s + k2 s + k1

Then a, b must satisfy the equation a(s + k1 ) + b(s + k2 ) = k1 m0 . This equation can be
rewritten as (a + b)s + (ak1 + bk2 ) = k1 m0 . Equating like coefficients, we have a + b = 0,
ak1 + bk2 = k1 m0 . Solving this system of equations, we have a = k1 m0 /(k1 − k2 ) and
b = k1 m0 /(k2 − k1 ). Thus

k1 m0 /(k1 − k2 ) k1 m0 /(k2 − k1 )
M2 = + ,
s + k2 s + k1
328 CHAPTER 5. THE LAPLACE TRANSFORM

which implies
   
k1 m0 −1 1 k1 m0 −1 1 k1 m0 −k2 t k1 m0 −k1 t
m2 (t) = L + L = e + e .
k1 − k2 s + k2 k2 − k1 s + k1 k1 − k2 k2 − k1

Finally, the equation for M3 becomes

k2 k1 m0
sM3 = k2 M2 = ,
(s + k2 )(s + k1 )

which implies
k2 k1 m0
M3 = .
s(s + k2 )(s + k1 )
Using partial fractions, we rewrite this equation as

k2 k1 m0 a b c
= + + .
s(s + k2 )(s + k1 ) s s + k2 s + k1

Then a, b, c must satisfy the equation a(s + k2 )(s + k1 ) + bs(s + k1 ) + cs(s + k2 ) = k2 k1 m0 .

This equation can be rewritten as (a + b + c)s2 + (ak1 + ak2 + bk1 + ck2 )s + ak1 k2 = k1 k2 m0 .
Equating like coefficients, we have a+b+c = 0, ak1 +ak2 +bk1 +ck2 = 0 and ak1 k2 = k1 k2 m0 .
The solution of this system is given by a = m0 , b = −m0 k1 /(k1 − k2 ) and c = m0 k2 /(k1 − k2 ).
Thus
m0 −m0 k1 /(k1 − k2 ) m0 k2 /(k1 − k2 )
M3 = + + ,
s s + k2 s + k1
and then
     
1 −1 m0 k1 −1 1 m0 k2 −1 1
m3 (t) = m0 L − L + L
s k1 − k2 s + k2 k1 − k2 s + k1
m0 k1 −k2 t m0 k2 −k1 t
= m0 − e + e .
k1 − k2 k1 − k2

5.5 Discontinuous Functions and Periodic Functions

1.
5.5. DISCONTINUOUS FUNCTIONS AND PERIODIC FUNCTIONS 329

2.

3.

4.
330 CHAPTER 5. THE LAPLACE TRANSFORM

5.

6.

7. Using the Heaviside function, we can write f (t) = (t − 9)5 u9 (t). The Laplace transform
has the property that L(uc (t)f (t − c)) = e−cs L(f (t)). Therefore,
120e−9s
L(u9 (t)(t − 9)5 ) = e−9s L(t5 ) = .
s6

8. Using the Heaviside function, we can write f (t) = u3 (t)[(t − 3)2 + 9]. The Laplace
transform has the property that L(uc (t)f (t − c)) = e−cs L(f (t)). Therefore,
 
2 −3s 2 −3s 2 9
L(u3 (t)[(t − 3) + 9]) = e L(t + 9) = e + .
s3 s

9. Using the Heaviside function, we can write f (t) = (t − π)[uπ (t) − u2π (t)]. The Laplace
transform has the property that L(uc (t)f (t − c)) = e−cs L(f (t)). Therefore,
L((t − π)[uπ (t) − u2π (t)]) = L(uπ (t)(t − π)) − L(u2π (t)(t − 2π)) − πL(u2π (t))
e−πs e−2πs πe−2πs
= e−πs L(t) − e−2πs L(t) − πe−2πs L(1) = 2 − 2 − .
s s s

10. The Laplace transform has the property that L(uc (t)f (t − c)) = e−cs L(f (t)). Therefore,
e−s + 3e−5s − 4e−8s
L(u1 +3u5 −4u8 ) = L(u1 )+3L(u5 )−4L(u8 ) = [e−s +3e−5s −4e−8s ]L(1) = .
s
5.5. DISCONTINUOUS FUNCTIONS AND PERIODIC FUNCTIONS 331

11. The Laplace transform has the property that L(uc (t)f (t − c)) = e−cs L(f (t)). Therefore,

L((t − 4)u3 (t) − (t − 3)u4 (t)) = L((t − 3)u3 ) − L(u3 ) − L((t − 4)u4 ) − L(u4 )
   
−3s −3s −4s −4s −3s 1 1 −4s 1 1
= e L(t) − e L(1) − e L(t) − e L(1) = e − −e + .
s2 s s2 s

12. The Laplace transform has the property that L(uc (t)f (t − c)) = e−cs L(f (t)). Therefore,
1 − e−s
L(t − (t − 1)u1 (t)) = L(t) − L((t − 1)u1 (t)) = .
s2

13. Using the fact that L−1 (e−cs G(s)) = uc (t)g(t − c), we see that
5!e−s
 
−1
L = u1 (t)g(t − 1),
(s − 5)6
where  
−1 5!
g(t) = L = t5 e5t .
(s − 5)6
Therefore,
L−1 (F (s)) = u1 (t)(t − 1)5 e5(t−1) .

14. Using the fact that L−1 (e−cs G(s)) = uc (t)g(t − c), we see that
e−2s
 
−1
L = u2 (t)g(t − 2),
s2 + s − 2
where  
−1 1
g(t) = L 2
.
s +s−2
Using partial fractions, we write
 
1 1 1 1
= − .
s2 + s − 2 3 s−1 s+2
Therefore,
1
g(t) = [et − e−2t ].
3
Thus
1
L−1 (F (s)) = u2 (t)[et−2 − e−2(t−2) ].
3
15. Using the fact that L−1 (e−cs G(s)) = uc (t)g(t − c), we see that
 −2s 
−1 e 2(s − 1)
L = u2 (t)g(t − 2),
s2 − 2s + 2
where  
−1 2(s − 1)
g(t) = L 2
.
s − 2s + 2
332 CHAPTER 5. THE LAPLACE TRANSFORM

Completing the square in the denominator, we write

2(s − 1) 2(s − 1)
= .
s2− 2s + 2 (s − 1)2 + 1

Therefore, g(t) = 2et cos t, and then

L−1 (F (s)) = 2u2 (t)et−2 cos(t − 2).

16. Using the fact that L−1 (e−cs G(s)) = uc (t)g(t − c), we see that
 −7s 
−1 8e
L = u7 (t)g(t − 7),
s2 − 64

where  
−1 8
g(t) = L 2
.
s − 64
Using partial fractions, we write
 
8 1 1 1
= − .
s2 − 64 2 s−8 s+8

Therefore,
1
g(t) = [e8t − e−8t ],
2
and then
1
L−1 (F (s)) = u7 (t)[e8(t−7) − e−8(t−7) ].
2
17. Using the fact that L−1 (e−cs G(s)) = uc (t)g(t − c), we see that

(s − 2)e−s
 
−1
L = u1 (t)g(t − 1),
s2 − 4s + 3

where  
−1 s−2
g(t) = L .
s2 − 4s + 3
Using partial fractions, we write
 
s−2 1 1 1
= + .
s2 − 4s + 3 2 s−1 s−3

Therefore,
1
g(t) = [et + e3t ],
2
and then
1
L−1 (F (s)) = [et−1 + e3(t−1) ]u1 (t).
2
5.5. DISCONTINUOUS FUNCTIONS AND PERIODIC FUNCTIONS 333

18. Using the fact that L−1 (e−cs G(s)) = uc (t)g(t − c), we see that
 −s
e + e−2s − e−3s − e−4s

−1
L = u1 (t)g(t − 1) + u2 (t)g(t − 2) − u3 (t)g(t − 3) − u4 (t)g(t − 4),
s
where  
−1 1
g(t) = L = 1.
s
Therefore,
L−1 (F (s)) = u1 (t) + u2 (t) − u3 (t) − u4 (t).

19. By the definition of the Laplace transform (for s > 0),

Z ∞ Z 1
−st 1 − e−s
L(f (t)) = f (t)e dt = e−st dt = .
0 0 s

20. Using the definition of the Laplace transform, we obtain

Z ∞ Z 1 Z 3
−st −st
L(f (t)) = f (t)e dt = e dt + e−st dt
0 0 2
−s −2s −3s −s
1−e e −e 1−e + e−2s − e−3s
= + = .
s s s

21. Using the definition of the Laplace transform, we have (for s > 0)
Z ∞ Z 1 Z 3 Z 5 Z 2n+1
−st −st −st −st
L(f (t)) = f (t)e dt = e dt + e dt + e dt + . . . + e−st dt
0 0 2 4 2n
 1−e −(2n+2)s
1
= 1 − e−s + e−2s − e−3s + . . . + e−2ns − e−(2n+1)s =
s s(1 + e−s )

22. Using Theorem 5.5.3, we know that

FT (s)
L(f (t)) = ,
1 − e−sT
where T is the period and Z T
FT (s) = e−st f (t) dt.
0
Here, T = 2. Therefore,
Z 2 Z 1 Z 2
−st −st
FT (s) = e f (t) dt = e dt − e−st dt
0 0 1
−s −s −2s −s
1−e e −e 1 − 2e + e−2s
= − = .
s s s
Thus (for s > 0)
1 − 2e−s + e−2s (1 − e−s )2 1 − e−s
L(f (t)) = = = .
s(1 − e−2s ) s(1 − e−s )(1 + e−s ) s(1 + e−s )
334 CHAPTER 5. THE LAPLACE TRANSFORM

23. Using Theorem 5.5.3, we know that

FT (s)
L(f (t)) = ,
1 − e−sT
where T is the period and Z T
FT (s) = e−st f (t) dt.
0
Here, T = 1. Therefore,
1 1
1 − se−s − e−s
Z Z
−st
FT (s) = e f (t) dt = e−st t dt = .
0 0 s2

Thus (for s > 0)

1 − se−s − e−s
L(f (t)) = .
s2 (1 − e−s )
24. Using Theorem 5.5.3, we know that
FT (s)
L(f (t)) = ,
1 − e−sT
where T is the period and Z T
FT (s) = e−st f (t) dt.
0
Here, T = π. Therefore,
π π
1 + e−πs
Z Z
−st
FT (s) = e f (t) dt = e−st sin(t) dt = .
0 0 s2 + 1

Thus (for s > 0)

1 + e−πs
L(f (t)) = .
(s2 + 1)(1 − e−πs )

25.(a) For s > 0:

1 e−s
L(f (t)) = L(1) − L(u1 (t)) = − .
s s
5.5. DISCONTINUOUS FUNCTIONS AND PERIODIC FUNCTIONS 335

(b)

Let F (s) = L[1 − u1 (t)]. Then

Z t
1 − e−s

F (s)
L [1 − u1 (ξ)] dξ = = .
0 s s2

(c)

Let G(s) = L(g(t)). Then

1 − e−s −s 1 − e
−s
(1 − e−s )2
L(h(t)) = G(s) − e−s G(s) = − e = .
s2 s2 s2

26.(a)
336 CHAPTER 5. THE LAPLACE TRANSFORM

(b) The given function is periodic. By Theorem 5.5.3,

FT (s)
L(p(t)) = ,
1 − e−sT
where T is the period and Z T
FT (s) = e−st f (t) dt.
0
Here the period is T = 2 and by problem 25(c),
(1 − e−s )2
F2 (s) = .
s2
Therefore,
(1 − e−s )2 (1 − e−s )2 1 − e−s
L(p(t)) = = = .
s2 (1 − e−2s ) s2 (1 − e−s )(1 + e−s ) s2 (1 + e−s )
(c) By Theorem 5.2.2,
L(p0 (t)) = sL(p(t)) − p(0).
Further, p0 (t) = f (t) where f is the periodic square wave in problem 22. Therefore,
1 − e−s
L(p0 (t)) = L(f (t)) = .
s(1 + e−s )
We also note that p(0) = 0. Therefore,
1 1 − e−s
L(p(t)) = L(p0 (t)) = 2 .
s s (1 + e−s )

5.6 Differential Equations with Discontinuous Forcing

Functions
1. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 )+L(y) =
L(f (t)), which is [s2 Y (s) − sy(0) − y 0 (0)] + Y (s) = L(f (t)). Applying the initial conditions,
we get s2 Y (s) − 5s − 3 + Y (s) = L(f (t)). The forcing function f (t) can be written as
f (t) = u0 (t) − uπ/2 (t). Therefore,

1 − e−πs/2
L(f (t)) = L(u0 (t)) − L(uπ/2 (t)) = .
s
Thus the equation for Y becomes

1 − e−πs/2
[s2 + 1]Y (s) = 5s + 3 + ,
s
and then
5s + 3 1 − e−πs/2
Y (s) = + .
s2 + 1 s(s2 + 1)
5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 337

Using partial fractions, we write the second term as

 
−πs/2 1 s
(1 − e ) − .
s s2 + 1

We obtain that
4s + 3 1 e−πs/2 se−πs/2
Y (s) = 2 + − + 2 .
s +1 s s s +1
Then, using the fact that L−1 (e−cs G(s)) = uc (t)g(t − c), we conclude that

y(t) = 4 cos t + 3 sin t + 1 − uπ/2 (t) + uπ/2 (t) cos(t − π/2).

2. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 ) +
2L(y 0 )+2L(y) = L(h(t)), which is [s2 Y (s)−sy(0)−y 0 (0)]+2[sY (s)−y(0)]+2Y (s) = L(h(t)).
Applying the initial conditions, we get s2 Y (s) − 5s − 4 + 2sY (s) − 10 + 2Y (s) = L(h(t)).
The forcing function h(t) can be written as h(t) = uπ (t) − u2π (t). Therefore,

e−πs − e−2πs
L(h(t)) = L(uπ (t)) − L(u2π (t)) = .
s
Thus the equation for Y becomes

2 e−πs − e−2πs
[s + 2s + 2]Y (s) = 5s + 14 + ,
s
and then
5s + 14 e−πs − e−2πs
Y (s) = 2 + .
s + 2s + 2 s(s2 + 2s + 2)
Using partial fractions, we write the second term as
 
1 −πs −2πs 1 2+s
(e −e ) − .
2 s s2 + 2s + 2

We obtain that
 
5s + 14 1 −πs −2πs 1 2+s
Y (s) = 2 + (e −e ) − .
s + 2s + 2 2 s s2 + 2s + 2
338 CHAPTER 5. THE LAPLACE TRANSFORM

Completing the square in the denominators, we conclude that

5(s + 1) + 9 e−πs 1 e−2πs 1
   
(s + 1) + 1 (s + 1) + 1
Y (s) = + − − − .
(s + 1)2 + 1 2 s (s + 1)2 + 1 2 s (s + 1)2 + 1

Then, using the fact that L−1 (e−cs G(s)) = uc (t)g(t − c), we conclude that
1
y(t) = 5e−t cos t + 9e−t sin t + uπ (t)(1 − e−(t−π) cos(t − π) − e−(t−π) sin(t − π))
2
1
− u2π (t)(1 − e−(t−2π) cos(t − 2π) − e−(t−2π) sin(t − 2π)).
2

The solution starts out as a free decaying oscillation. The amplitude increases while the
forcing is present. After the forcing is removed, the solution decays due to the damping
term.
3. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 ) +
4L(y) = L(sin(t) − u2π (t) sin(t − 2π)), which is

1 − e−2πs
[s2 Y (s) − sy(0) − y 0 (0)] + 4Y (s) = .
s2 + 1
Applying the initial conditions, we get
1 − e−2πs
s2 Y (s) + 4Y (s) = .
s2 + 1
Therefore, the equation for Y becomes

2 1 − e−2πs
[s + 4]Y (s) = 2 ,
s +1
thus
1 − e−2πs
Y (s) = 2 .
(s + 4)(s2 + 1)
Using partial fractions, we can write
 
1 −2πs 1 1
Y (s) = (1 − e ) 2 − .
3 s + 1 s2 + 4
5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 339

Therefore, we conclude that

 
1 1 1
y(t) = sin t − sin 2t − u2π (t) sin(t − 2π) + u2π (t) sin(2(t − 2π)) .
3 2 2

4. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 ) +
4L(y) = L(sin(t) + uπ (t) sin(t − π)), which is

1 + e−πs
[s2 Y (s) − sy(0) − y 0 (0)] + 4Y (s) = .
s2 + 1

Applying the initial conditions, we get

1 + e−πs
s2 Y (s) + 4Y (s) = .
s2 + 1

Therefore, the equation for Y becomes

1 + e−πs
[s2 + 4]Y (s) = ,
s2 + 1

thus
1 + e−πs
Y (s) = .
(s2 + 4)(s2 + 1)
Using partial fractions, we can write
 
1 −πs 1 1
Y (s) = (1 + e ) 2 − .
3 s + 1 s2 + 4

Therefore, we conclude that

 
1 1 1
y(t) = sin t + uπ (t) sin(t − π) − sin 2t − uπ (t) sin(2(t − π)) .
3 2 2
340 CHAPTER 5. THE LAPLACE TRANSFORM

Since there is no damping term, the solution follows the forcing function, after which the
solution is a steady oscillation about y = 0.
5. First, we write the nonhomogeneous term as f (t) = u0 (t) − u10 (t). Then, let Y (s) = L(y).
Applying the Laplace transform, we have L(y 00 ) + 3L(y 0 ) + 2L(y) = L(u0 (t) − u10 (t)), which
is
2 0 1 − e−10s
[s Y (s) − sy(0) − y (0)] + 3[sY (s) − y(0)] + 2Y (s) = .
s
Applying the initial conditions, we get
1 − e−10s
s2 Y (s) + 3sY (s) + 2Y (s) = .
s
Therefore, the equation for Y becomes
1 − e−10s
[s2 + 3s + 2]Y (s) = ,
s
thus
1 − e−10s
Y (s) = .
s(s2 + 3s + 2)
Using partial fractions, we can write
 
1 −10s 1 1 2
Y (s) = (1 − e ) + − .
2 s s+2 s+1
Therefore, we conclude that
1 1
y(t) = [1 + e−2t − 2e−t ] − u10 (t)[1 + e−2(t−10) − 2e−(t−10) ].
2 2
5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 341

The solution increases to a temporary steady value of y = 1/2. After the forcing ceases, the
response decays exponentially to y = 0.
6. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 ) +
3L(y 0 ) + 2L(y) = L(u2 (t)), which is

e−2s
[s2 Y (s) − sy(0) − y 0 (0)] + 3[sY (s) − y(0)] + 2Y (s) = .
s
Applying the initial conditions, we have

2 e−2s
s Y (s) − 6s − 6 + 3sY (s) − 18 + 2Y (s) = .
s
Therefore, the equation for Y becomes

e−2s
[s2 + 3s + 2]Y (s) = 6s + 24 + ,
s
thus
6s + 24 e−2s
Y (s) = + .
s2 + 3s + 2 s(s2 + 3s + 2)
Using partial fractions, we can write

e−2s 1
 
18 12 1 2
Y (s) = − + + − .
s+1 s+2 2 s s+2 s+1

Therefore, we conclude that

1
y(t) = 18e−t − 12e−2t + u2 (t)[1 + e−2(t−2) − 2e−(t−2) ].
2

Due to the initial conditions, the response has a transient overshoot, followed by an expo-
nential convergence to a steady value of ys = 1/2.
7. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 )+L(y) =
L(u3π (t)), which is
e−3πs
[s2 Y (s) − sy(0) − y 0 (0)] + Y (s) = .
s
342 CHAPTER 5. THE LAPLACE TRANSFORM

Applying the initial conditions, we get

e−3πs
s2 Y (s) − s + Y (s) = .
s
Therefore, the equation for Y becomes
e−3πs
[s2 + 1]Y (s) = s + ,
s
thus
s e−3πs
Y (s) = + .
s2 + 1 s(s2 + 1)
Using partial fractions, we can write
 
s −3πs 1 s
Y (s) = 2 +e − .
s +1 s s2 + 1
Therefore, we conclude that

y(t) = cos t + u3π (t)[1 − cos(t − 3π)] = cos t + u3π (t)[1 + cos t].

Due to the initial conditions, the solution temporarily oscillated about y = 0. After the
forcing is applied, the response is a steady oscillation about ym = 1.
8. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 ) +
5
L(y 0 ) + L(y) = L(t − uπ/2 (t)(t − π/2)), which is
4
5 1 − e−πs/2
[s2 Y (s) − sy(0) − y 0 (0)] + [sY (s) − y(0)] + Y (s) = .
4 s2
Applying the initial conditions, we get
5 1 − e−πs/2
s2 Y (s) + sY (s) + Y (s) = .
4 s2
Therefore, the equation for Y becomes
1 − e−πs/2
[s2 + s + 5/4]Y (s) = ,
s2
5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 343

thus
1 − e−πs/2
Y (s) = 2 2 .
s (s + s + 5/4)
Using partial fractions, we can write
 
−πs/2 64s − 16 4 16
Y (s) = (1 − e ) + − .
25(4s2 + 4s + 5) 5s2 25s

Now, completing the square in the denominator of the first term on the right-hand side
above, we have
 
64s − 16 64s − 16 16 s + 1/2 3/4
= = − .
25(4s2 + 4s + 5) 100((s + 1/2)2 + 1) 25 (s + 1/2)2 + 1 (s + 1/2)2 + 1

This implies that

 
−πs/2 16 s + 1/2 3/4 5 1
Y (s) = (1 − e )· − + − .
25 (s + 1/2)2 + 1 (s + 1/2)2 + 1 4s2 s

Therefore, we conclude that

 
16 −t/2 3 −t/2 5
y(t) = e cos t − e sin t + t − 1
25 4 4
 
16 −(t−π/2)/2 3 −(t−π/2)/2 5
− uπ/2 (t) e cos(t − π/2) − e sin(t − π/2) + (t − π/2) − 1 .
25 4 4

The forcing term causes the initial oscillation, but then the solution approaches the steady
state ys (t) = 2π/5.
9. First, we can write the nonhomogeneous term as g(t) = t/2 − u6 (t)(t − 6)/2. Then, let
Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 ) + L(y) =
1
L(t − u6 (t)(t − 6)), which is
2
1 1 − e−6s
 
2 0
[s Y (s) − sy(0) − y (0)] + Y (s) = .
2 s2
344 CHAPTER 5. THE LAPLACE TRANSFORM

Applying the initial conditions, we have

1 1 − e−6s
s2 Y (s) − 6s − 8 + Y (s) = .
2 s2
Therefore, the equation for Y becomes

1 1 − e−6s
[s2 + 1]Y (s) = 6s + 8 + ,
2 s2
thus
6s + 8 1 1 − e−6s
Y (s) = + .
s2 + 1 2 s2 (s2 + 1)
Using partial fractions, we can write
 
s 1 1 −6s 1 1
Y (s) = 6 2 +8 2 + (1 − e ) 2 − 2 .
s +1 s +1 2 s s +1

Therefore, we conclude that

1
y(t) = 6 cos t + 8 sin t + [t − sin t − u6 (t)(t − 6) + u6 (t) sin(t − 6)] .
2

The solution increases in response to the ramp input, and, thereafter oscillated about a mean
value of ym = 3.
10. First, we can write the forcing term as g(t) = (1 − uπ (t)) sin t = sin t + uπ (t) sin(t − π).
Then, let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 ) +
5
L(y 0 ) + L(y) = L(sin(t) + uπ (t) sin(t − π)), which is
4

2 0 5 1 + e−πs
[s Y (s) − sy(0) − y (0)] + [sY (s) − y(0)] + Y (s) = 2 .
4 s +1
Applying the initial conditions, we get

5 1 + e−πs
s2 Y (s) + sY (s) + Y (s) = 2 .
4 s +1
5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 345

Therefore, the equation for Y becomes

1 + e−πs
[s2 + s + 5/4]Y (s) = ,
s2 + 1
thus
1 + e−πs
Y (s) = .
(s2 + s + 5/4)(s2 + 1)
Using partial fractions, we can write
 
1 −πs 16(s + 1/2) + 4 16s 4
Y (s) = (1 + e ) − 2 + .
17 (s + 1/2)2 + 1 s + 1 s2 + 1

Therefore, we conclude that

1  −t/2
cos t + 4e−t/2 sin t − 16 cos t + 4 sin t

y(t) = 16e
17
1
uπ (t) 16e−(t−π)/2 cos(t − π) + 4e−(t−π)/2 sin(t − π) − 16 cos(t − π) + 4 sin(t − π) .
 
+
17

The initial oscillation is a response to the forcing term, but after the forcing term is taken
away, the solution decays to zero due to the damping term.
11. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 ) +
4L(y) = L(uπ (t) − u3π (t)), which is

e−πs − e−3πs
[s2 Y (s) − sy(0) − y 0 (0)] + 4Y (s) = .
s
Applying the initial conditions, we get

e−πs − e−3πs
s2 Y (s) − 3s − 7 + 4Y (s) = .
s
Therefore, the equation for Y becomes

e−πs − e−3πs
[s2 + 4]Y (s) = 3s + 7 + ,
s
346 CHAPTER 5. THE LAPLACE TRANSFORM

thus
3s + 7 e−πs − e−3πs
Y (s) = + .
s2 + 4 s(s2 + 4)
Using partial fractions, we can write
 
s 7 2 −πs −3πs 1 1 s
Y (s) = 3 2 + + (e −e ) − .
s + 4 2 s2 + 4 4 s s2 + 4
Therefore, we conclude that
7 1 1
y(t) = 3 cos 2t + sin 2t + [1 − cos(2(t − π))]uπ (t) − [1 − cos(2(t − 3π))]u3π (t).
2 4 4

Since there is no damping term, the solution responds immediately to the forcing term.
There is a temporary oscillation about y = 1/4.
12. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 0000 ) −
L(y) = L(u1 (t) − u2 (t)), which is

e−s − e−2s
[s4 Y (s) − s3 y(0) − s2 y 0 (0) − sy 00 (0) − y 000 (0)] − Y (s) = .
s
Applying the initial conditions, we get

e−s − e−2s
s4 Y (s) − 12s3 − 7s2 − 2s + 9 − Y (s) = .
s
Therefore, the equation for Y becomes

e−s − e−2s
[s4 − 1]Y (s) = 12s3 + 7s2 + 2s − 9 + ,
s
thus
12s3 + 7s2 + 2s − 9 e−s − e−2s
Y (s) = + .
s4 − 1 s(s4 − 1)
Using partial fractions, we can write
 
3 4 5s + 8 −s −2s 1 −4 1 1 2s
Y (s) = + + + (e − e ) + + + .
s − 1 s + 1 s2 + 1 4 s s + 1 s − 1 s2 + 1
5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 347

Therefore, we conclude that

y(t) = 3et + 4e−t + 5 cos t + 8 sin t − [u1 (t) − u2 (t)]

1 −(t−1) 1
+ [e + et−1 + 2 cos(t − 1)]u1 (t) − [e−(t−2) + et−2 + 2 cos(t − 2)]u2 (t).
4 4

The solution increases exponentially.

13. Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 0000 ) +
5L(y 00 ) + 4L(y) = L(1 − uπ (t)), which is

1 − e−πs
[s4 Y (s) − s3 y(0) − s2 y 0 (0) − sy 00 (0) − y 000 (0)] + 5[s2 Y (s) − sy(0) − y 0 (0)] + 4Y (s) = .
s
Applying the initial conditions, we get

1 − e−πs
s4 Y (s) + 5s2 Y (s) + 4Y (s) = .
s
Therefore, the equation for Y becomes

1 − e−πs
[s4 + 5s2 + 4]Y (s) = ,
s
thus
1 − e−πs
Y (s) = .
s(s4 + 5s2 + 4)
Using partial fractions, we can write
 
−πs 1 3 s 4s
Y (s) = (1 − e ) + − .
12 s s2 + 4 s2 + 1

Therefore, we conclude that

1 1
y(t) = [3 + cos 2t − 4 cos t] − uπ (t) [3 + cos(2(t − π)) − 4 cos(t − π)].
12 12
348 CHAPTER 5. THE LAPLACE TRANSFORM

After an initial transient, the solution oscillates about ym = 0.

14. The specified function is defined by
0 ≤ t < t0

 0
h

f (t) = (t − t0 ) t0 ≤ t < t0 + k
 k

h t0 + k ≤ t.
Using uc , this function can be written as
h h
f (t) = (t − t0 )ut0 (t) − (t − t0 − k)ut0 +k (t).
k k
15. The specified function is defined by


 0 0 ≤ t < t0

 h
(t − t0 ) t0 ≤ t < t0 + k


f (t) = k
h

 − (t − t0 − 2k) t0 + k ≤ t < t0 + 2kk
 k



h t0 + 2k ≤ t.
Using uc , this function can be written as
h 2h h
f (t) = (t − t0 )ut0 (t) − (t − t0 − k)ut0 +k (t) + (t − t0 − 2k)ut0 +2k (t).
k k k

16.(a) The plot shows k = 2.

5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 349

1
(b) Let U (s) = L(u). Applying the Laplace transform, we have L(u00 ) + L(u0 ) + L(y) =
4
kL(g(t)), which is

1 e−3s/2 − e−5s/2
[s2 U (s) − su(0) − u0 (0)] + [sU (s) − u(0)] + U (s) = k .
4 s
Applying the initial conditions, we get

1 e−3s/2 − e−5s/2
s2 U (s) + sU (s) + U (s) = k .
4 s
Therefore, the equation for U becomes

e−3s/2 − e−5s/2
[s2 + 1/4s + 1]U (s) = k ,
s
thus
e−3s/2 − e−5s/2
U (s) = k .
s(s2 + 1/4s + 1)
Using partial fractions, we can write
   
−3s/2 −5s/2 1 4s + 1 −3s/2 −5s/2 1 (s + 1/8) + 1/8
U (s) = k(e −e ) − = k(e −e ) − .
s 4s2 + s + 4 s (s + 1/8)2 + 63/64
Therefore, we conclude that
" √ ! √ !#
63(t − 3/2) e−(t−3/2)/8 63(t − 3/2)
u(t) = ku 3 (t) 1 − e−(t−3/2)/8 cos − √ sin
2 8 63 8
" √ ! √ !#
63(t − 5/2) e−(t−5/2)/8 63(t − 5/2)
− ku 5 (t) 1 − e−(t−5/2)/8 cos − √ sin .
2 8 63 8

(c)

(d) Due to the damping term, the solution will decay to zero. The maximum will occur
shortly after the forcing ceases. By plotting the various solutions, it appears that the solution
will reach a value of u = 2 as long as k > 2.51.
350 CHAPTER 5. THE LAPLACE TRANSFORM

(e) Based on the graph and numerical calculation, |u(t)| < 0.1 for t > 25.6773.
17. In this problem, we consider the initial value problem
1
y 00 + 4y = [(t − 5)u5 (t) − (t − 5 − k)u5+k (t)].
k
(a) The plot shows k = 5.

The problem is identical to Example 1 when k = 5.

(b) Let Y (s) = L(y). Applying the Laplace transform to the equation, we have L(y 00 ) +
4L(y) = L(f (t)), which is
e−5s e−(5+k)s
[s2 Y (s) − sy(0) − y 0 (0)] + 4Y (s) = − .
ks2 ks2
Applying the initial conditions, we have
e−5s e−(5+k)s
s2 Y (s) + 4Y (s) = − .
ks2 ks2
Therefore, the equation for Y becomes
e−5s e−(5+k)s
Y (s) = − .
ks2 (s2 + 4) ks2 (s2 + 4)
Using partial fractions, we can write
 
−5s −(5+k)s 1 1 1
Y (s) = (e −e ) − .
4k s2 s2 + 4
Thus we conclude that
   
u5 (t) 1 u5+k (t) 1
y(t) = (t − 5) − sin(2(t − 5)) − (t − 5 − k) + sin(2(t − 5 − k)) .
4k 2 4k 2

(c) We note that for t > 5 + k, the solution is given by

1 1 1 1 sin k
y(t) = − sin(2t − 10) + sin(2t − 10 − 2k) = − cos(2t − 10 − k).
4 8k 8k 4 4k
Therefore, for t > 5 + k, the solution oscillates about ym = 1/4 with an amplitude of
A = | sin k|/4k. The figures show the cases k = 3, k = 4, and k = 5, respectively.
5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 351

18.(a) The plot shows k = 2.

1
(b) We can write fk (t) = [u4−k (t) − u4+k (t)]. Let Y (s) = L(y). Applying the Laplace
2k
transform to the equation, we have
1 1
L(y 00 ) + L(y 0 ) + 4L(y) = L(u4−k (t) − u4+k (t)),
3 2k
which is
1 e−(4−k)s e−(4+k)s
[s2 Y (s) − sy(0) − y 0 (0)] + [sY (s) − y(0)] + 4Y (s) = − .
3 2ks 2ks
Applying the initial conditions, we get

1 e−(4−k)s e−(4+k)s
s2 Y (s) + sY (s) + 4Y (s) = − .
3 2ks 2ks
Therefore, the equation for Y becomes

3e−(4−k)s 3e−(4+k)s
Y (s) = − .
2ks(3s2 + s + 12) 2ks(3s2 + s + 12)
Using partial fractions, we can write
   
1 1 1 1 1 1 1 1 + 6(s + 1/6)
= − = − .
s(3s2 + s + 12) 12 s 3s2 + s + 12 12 s 6 (s + 1/6)2 + 143/36
352 CHAPTER 5. THE LAPLACE TRANSFORM

Let  
1 1 1/6 s + 1/6
H(s) = − − .
8k s (s + 1/6)2 + 143/36 (s + 1/6)2 + 143/36
Therefore, letting
−t/6
" √ ! √ !#
1 e 1 143 t 143 t
h(t) = L−1 (H(s)) = − √ sin + cos ,
8k 8k 143 6 6

we conclude that
y(t) = h(t − 4 + k)u4−k (t) − h(t − 4 − k)u4+k (t).

(c)

As k decreases, the solution remains null for a longer period of time. In addition, the
amplitude increases.
19.(a) Let Y (s) = L(y). Applying the Laplace transform to the equation, we have
n
X
n
1+2 (−1)k e−kπs
X k=1
[s2 Y (s) − sy(0) − y 0 (0)] + Y (s) = L(1 + 2 (−1)k ukπ (t)) = .
k=1
s

Applying the initial conditions, we have

n
X
1+2 (−1)k e−kπs
k=1
[s2 + 1]Y (s) = .
s
Therefore,
n
X
1+2 (−1)k e−kπs
k=1
Y (s) = .
s(s2 + 1)
Using partial fractions, we see that
n
! 
X
k −kπs 1 s
Y (s) = 1+2 (−1) e − 2 .
k=1
s s + 1
5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 353

Thus we conclude that

n
X n
X
y(t) = 1 + 2 (−1)k ukπ (t) − cos(t) − 2 (−1)k ukπ (t) cos(t − kπ)
k=1 k=1
n
X
= 1 − cos(t) + 2 (−1)k [1 − cos(t − kπ)]ukπ (t).
k=1

(b)

(c) As n → ∞, the amplitude of the solution increases. Below is a graph for the case n = 30.

20.(a) Let Y (s) = L(y). Applying the Laplace transform to the equation, we have

[s2 Y (s) − sy(0) − y 0 (0)] + 0.1[sY (s) − y(0)] + Y (s)

Xn

n
1+2 (−1)k e−kπs
X k=1
= L(1 + 2 (−1)k ukπ (t)) = .
k=1
s

Applying the initial conditions, we get

n
X
1+2 (−1)k e−kπs
k=1
[s2 + 0.1s + 1]Y (s) = .
s
354 CHAPTER 5. THE LAPLACE TRANSFORM

Therefore,
n
X
1+2 (−1)k e−kπs
k=1
Y (s) = .
s(s2 + 0.1s + 1)
Using partial fractions, we see that
n
! 
X
k −kπs 1 s + 0.1
Y (s) = 1+2 (−1) e − 2
k=1
s s + 0.1s + 1
n
!  
X
k −kπs 1 s + 0.05 0.05
= 1+2 (−1) e − 2 + 0.9975
− 2 + 0.9975
.
k=1
s (s + 0.05) (s + 0.05)
Therefore, we conclude that
n
X
y(t) = h(t) + 2 (−1)k ukπ (t)h(t − kπ),
k=1

where
√ 0.05 −0.05t √
h(t) = 1 − e−0.05t cos( 0.9975 t) − √ e sin( 0.9975 t).
0.9975
(b) The plots show n = 10, n = 15, and n = 20, respectively.

√will approach y = −1 while for n even,

(c) Due to the damping term, for n odd, the solution
the solution will approach y = 1. The frequency is 0.9975 ≈ 1.
21.(a)
5.6. DIFFERENTIAL EQS. WITH DISCONTINUOUS FORCING FUNCTIONS 355

(b) Let Y (s) = L(y). Applying the Laplace transform to the equation, we have
n
X
1+ (−1)k e−kπs
k=1
[s2 Y (s) − sy(0) − y 0 (0)] + Y (s) = .
s
Applying the initial conditions, we get
n
X
1+ (−1)k e−kπs
k=1
[s2 + 1]Y (s) = .
s
Therefore,
n
X
1+ (−1)k e−kπs
k=1
Y (s) = .
s(s2 + 1)
Using partial fractions, we see that
n
! 
X
k −kπs 1 s
Y (s) = 1+ (−1) e − .
k=1
s s2 + 1

Thus we conclude that

n
X n
X
y(t) = 1 + (−1)k ukπ (t) − cos(t) − (−1)k ukπ (t) cos(t − kπ)
k=1 k=1
n
X
= 1 − cos(t) + (−1)k [1 − cos(t − kπ)]ukπ (t).
k=1

(c)

There is no damping there. Each interval of forcing adds energy to the system. Therefore,
the amplitude will continue to increase until increase. For n = 15, g(t) = 0 when t > 15π.
Therefore, the oscillation will eventually become steady.
356 CHAPTER 5. THE LAPLACE TRANSFORM

(d) If n is even, the forcing term g(t) = 1 for large values of t, while for n odd, the forcing
term g(t) = 0 for large values of t. Therefore, for n even, the solution will eventually
oscillate about the line y = 1, while for n odd, the solution will eventually oscillate about
the line y = 0. As n increases, the amount of forcing increases, causing the amplitude of the
oscillations to increase until they reach a steady state after t > nπ.
22.(a) The plots show n = 10, n = 15, and n = 20, respectively.

(b) For n odd, the solution approaches y = 0. For n even, the solution approaches y = 1.
The solution is a sum of damped sinusoids, each of frequency ω ≈ 1.
(c) In problem 20, the solution approaches y = 1 if n is even and y = −1 if n is odd.
23.(a) Let Y (s) = L(y). Applying the Laplace transform to the equation, we have
n
X
1+2 (−1)k e(−11k/4)s
k=1
[s2 Y (s) − sy(0) − y 0 (0)] + Y (s) = .
s

Applying the initial conditions, we get

n
X
1+2 (−1)k e(−11k/4)s
k=1
[s2 + 1]Y (s) = .
s

Therefore,
n
X
1+2 (−1)k e(−11k/4)s
k=1
Y (s) = .
s(s2 + 1)
Using partial fractions, we see that

n
! 
X 1 s
Y (s) = 1+2 (−1)k e(−11k/4)s − .
k=1
s s2 + 1
5.7. IMPULSE FUNCTIONS 357

Thus we conclude that

n
X n
X
k
y(t) = 1 + 2 (−1) u11k/4 (t) − cos(t) − 2 (−1)k u11k/4 (t) cos(t − 11k/4)
k=1 k=1
n
X
= 1 − cos(t) + 2 (−1)k [1 − cos(t − 11k/4)]u11k/4 (t).
k=1

(b)

(c) Based on the plot, the “slow period” appears to be approximately 88 and the “fast
period” appears to be about 6.
(d) The natural frequency of the system is ω0 = 1. The forcing function is initially periodic
with period T = 11/2. Therefore, the corresponding forcing frequency is ω = 1.1424.
Therefore, the “slow frequency” is

|ω − ω0 |
ωs = = 0.0712
2
and the “fast frequency’ is
|ω + ω0 |
ωf = = 1.0712.
2
Based on these values, the slow period is predicted as 88.247 and the fast period is predicted
to be 5.8656.

5.7 Impulse Functions

1. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain

[s2 Y (s) − sy(0) − y 0 (0)] + 2[sY (s) − y(0)] + 2Y (s) = e−πs .

Applying the initial conditions, we get

[s2 Y (s) − 1] + 2sY (s) + 2Y (s) = e−πs ,

358 CHAPTER 5. THE LAPLACE TRANSFORM

which can be rewritten as

[s2 + 2s + 2]Y (s) − 1 = e−πs .
Therefore,
1 + e−πs 1 e−πs
Y (s) = = + .
s2 + 2s + 2 (s + 1)2 + 1 (s + 1)2 + 1
This implies that
y(t) = e−t sin t + uπ (t)e−(t−π) sin(t − π).

2. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain

[s2 Y (s) − sy(0) − y 0 (0)] + 4Y (s) = e−πs − e−2πs .

Applying the initial conditions, we get

[s2 + 4]Y (s) = e−πs − e−2πs .

Therefore,
e−πs − e−2πs
Y (s) = .
s2 + 4
This implies that
1 1 1
y(t) = sin(2t − 2π)uπ (t) − sin(2t − 4π)u2π (t) = [uπ (t) − u2π (t)] sin 2t.
2 2 2
5.7. IMPULSE FUNCTIONS 359

3. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain
e−10s
[s2 Y (s) − sy(0) − y 0 (0)] + 3[sY (s) − y(0)] + 2Y (s) = e−5s + .
s
Applying the initial conditions, we get
e−10s
[s2 Y (s) − 1/2] + 3sY (s) + 2Y (s) = e−5s + ,
s
which can be rewritten as
e−10s
[s2 + 3s + 2]Y (s) − 1/2 = e−5s + .
s
Therefore,
1/2 + e−5s e−10s
 −5s
e−5s
  −10s
e−10s e−10s

1 e e
Y (s) = 2 + = + − + − + .
s + 3s + 2 s(s2 + 3s + 2) 2 s+1 s+2 2(s + 2) s + 1 2s
This implies that
 
1 −t −2t −(t−5) −2(t−5) 1 −2(t−10) −(t−10) 1
y(t) = [e − e ] + u5 (t)[e −e ] + u10 (t) e −e + .
2 2 2

4. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain
[s2 Y (s) − sy(0) − y 0 (0)] − Y (s) = −20e−3s .
Applying the initial conditions, we get
[s2 Y (s) − 4s − 3] − Y (s) = −20e−3s ,
which can be rewritten as
[s2 − 1]Y (s) − 4s − 3 = −20e−3s .
Therefore,
4s + 3 20e−3s
Y (s) = − 2 .
s2 − 1 s −1
This implies that
y(t) = 4 cosh t + 3 sinh t − 20 sinh(t − 3)u3 (t).
360 CHAPTER 5. THE LAPLACE TRANSFORM

5. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain
1
[s2 Y (s) − sy(0) − y 0 (0)] + 2[sY (s) − y(0)] + 3Y (s) = + e−3πs .
s2 +1
Applying the initial conditions, we get
1
[s2 + 2s + 3]Y (s) = + e−3πs .
s2 +1
Therefore,
e−3πs e−3πs
 
1 1 −s + 1 s+1
Y (s) = 2 + = + + .
(s + 1)(s2 + 2s + 3) s2 + 2s + 3 4 s2 + 1 (s + 1)2 + 2 (s + 1)2 + 2
This implies that
√
1h −t
√ i 2 √
y(t) = − cos(t) + sin(t) + e cos( 2t) + u3π (t)e−(t−3π) sin( 2(t − 3π)).
4 2

6. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain

[s2 Y (s) − sy(0) − y 0 (0)] + 4Y (s) = e−4πs .

Applying the initial conditions, we get

[s2 Y (s) − s/2] + 4Y (s) = e−4πs ,

5.7. IMPULSE FUNCTIONS 361

which can be rewritten as

[s2 + 4]Y (s) − s/2 = e−4πs .
Therefore,
s e−4πs
Y (s) = + .
2(s2 + 4) s2 + 4
This implies that
1 1 1 1
y(t) = cos 2t + sin(2t − 8π)u4π (t) = cos 2t + u4π (t) sin 2t.
2 2 2 2

7. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain
[s2 Y (s) − sy(0) − y 0 (0)] + Y (s) = e−2πs .
Applying the initial conditions, we get
[s2 Y (s) − 1] + Y (s) = e−2πs ,
which can be rewritten as
[s2 + 1]Y (s) − 1 = e−2πs .
Therefore,
1 e−2πs
Y (s) = + .
s2 + 1 s2 + 1
This implies that
y(t) = sin t + u2π (t) sin(t − 2π) = [1 + u2π (t)] sin t.
362 CHAPTER 5. THE LAPLACE TRANSFORM

8. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain
[s2 Y (s) − sy(0) − y 0 (0)] + 4Y (s) = 2e−(π/4)s .
Applying the initial conditions, we get
[s2 + 4]Y (s) = 2e−(π/4)s .
Therefore,
2e−(π/4)s
Y (s) = .
s2 + 4
This implies that
y(t) = uπ/4 (t) sin(2(t − π/4)).

9. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain
e−(π/2)s e−2πs
[s2 Y (s) − sy(0) − y 0 (0)] + Y (s) = + 3e−(3π/2)s − .
s s
Applying the initial conditions, we get
e−(π/2)s e−2πs
[s2 + 1]Y (s) = + 3e−(3π/2)s − .
s s
Therefore,
e−(π/2)s 3e−(3π/2)s e−2πs
 −(π/2)s
se−(π/2)s 3e−(3π/2)s
 −2πs
se−2πs
 
e e
Y (s) = + 2 − 2 = − 2 + 2 − − 2 .
s(s2 + 1) s +1 s(s + 1) s s +1 s +1 s s +1
This implies that
y(t) = uπ/2 (t)[1 − cos(t − π/2)] + 3u3π/2 (t) sin(t − 3π/2) − u2π (t)[1 − cos(t − 2π)].
5.7. IMPULSE FUNCTIONS 363

10. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain
1
2[s2 Y (s) − sy(0) − y 0 (0)] + [sY (s) − y(0)] + 6Y (s) = e−(π/6)s .
2
Applying the initial conditions, we get
1
[2s2 + s + 6]Y (s) = e−(π/6)s .
2
Therefore,
e−(π/6)s
 
−(π/6)s 1/4
Y (s) = =e .
2(2s2 + s + 6) (s + 1/4)2 + (47/16)
This implies that
√ !
1 47
y(t) = uπ/6 (t) √ e−(t−π/6)/4 sin (t − π/6) .
47 4

11. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain
s
[s2 Y (s) − sy(0) − y 0 (0)] + 2[sY (s) − y(0)] + 2Y (s) = + e−(π/2)s .
s2 +1
Applying the initial conditions, we get
s
[s2 + 2s + 2]Y (s) = + e−(π/2)s .
s2 +1
Therefore,
s e−(π/2)s
Y (s) = +
(s2 + 1)(s2 + 2s + 2) s2 + 2s + 2
e−(π/2)s
 
1 s 2 s+1 3
= + − − + .
5 s2 + 1 s2 + 1 (s + 1)2 + 1 (s + 1)2 + 1 (s + 1)2 + 1
This implies that
1
cos t + 2 sin t − e−t cos t − 3e−t sin t + uπ/2 (t)e−(t−π/2) sin(t − π/2).

y(t) =
5
364 CHAPTER 5. THE LAPLACE TRANSFORM

12. Let Y (s) = L(y) and apply the Laplace transform to the differential equation. We obtain
[s4 Y (s) − s3 y(0) − s2 y 0 (0) − sy 00 (0) − y 000 (0)] − Y (s) = e−s .
Applying the initial conditions, we get
[s4 − 1]Y (s) = e−s .
Therefore,
e−s
 
−s 1 1 1
Y (s) = 4 =e − .
s −1 2 s2 − 1 s2 + 1
This implies that
1
y(t) = [sinh(t − 1) − sin(t − 1)]u1 (t).
2

13.(a) The equation will be of the form

2y 00 + y 0 + 2y = δ(t − 5) + kδ(t − t0 ).
We need to find k and t0 so that the system will rest again after exactly one cycle. Applying
the Laplace transform and using the initial conditions, this equation becomes
[2s2 + s + 2]Y (s) = e−5s + ke−t0 s .
Therefore,
e−5s + ke−t0 s
 
−5s −t0 s 1 1
Y (s) = = (e + ke ) .
2s2 + s + 2 2 (s + 1/4)2 + (15/16)
5.7. IMPULSE FUNCTIONS 365

This implies that

√ ! √ !
2 15 2 15
y(t) = u5 (t) √ e−(t−5)/4 sin (t − 5) + kut0 (t) √ e−(t−t0 )/4 sin (t − t0 ) .
15 4 15 4

In order for this system to return to equilibrium after one cycle, we need the magnitude of
these sinusoidal waves to be the same. Therefore, we need e5/4 = −ket0 /4 , and we need the
phase to differ by 2π. That is, we need
√ √
15 15
(t − 5) = (t − t0 ) + 2π.
4 4

8π
Solving this equation, we see that t0 = 5 + √ . Solving the equation above for k, we see
√ 15
−2π/ 15
that k = −e .
(b) From the analysis above, we see that the solution will be zero after time t0 .

14.(a) Applying the Laplace transform and using the initial conditions, the differential equa-
tion becomes
1
s2 Y (s) + sY (s) + Y (s) = e−s .
2
Solving for the transform, we get

e−s e−s
Y (s) = 2 = .
s + s/2 + 1 (s + 1/4)2 + (15/16)

Therefore,
√ !
4 15
y(t) = u1 (t) √ e−(t−1)/4 sin (t − 1) .
15 4
366 CHAPTER 5. THE LAPLACE TRANSFORM

(b) As shown on the graph, the maximum is attained at some t1 > 2. Using the equation
for y(t) and calculating y 0 (t), we see that t1 ≈ 2.3613. The maximum value is y(2.3613) ≈
0.71153.
(c) Setting γ = 1/4, the transform of the solution is

e−s
Y (s) = .
s2 + s/4 + 1

Using the same analysis as in part (a), we conclude that

√ !
8 3 7
y(t) = √ e−(t−1)/8 sin (t − 1) u1 (t).
3 7 8

By setting y 0 (t) = 0, we see that the maximum is attained at t1 ≈ 2.4569 with y(t1 ) ≈ 0.8335.
(d) Suppose that 0 < γ < 1. Then the transform of the solution is

e−s
Y (s) = ,
s2 + γs + 1

which implies

2 −γ(t−1)/2
p
2

y(t) = p e sin 1 − γ /4 · (t − 1) u1 (t).
4 − γ2
5.7. IMPULSE FUNCTIONS 367

Setting y 0 (t) = 0, we obtain

hp i 1p
tan 1 − γ 2 /4 · (t − 1) = 4 − γ 2.
γ

As γ → 0, we obtain the formal equation tan(t − 1) = ∞. Hence t1 → 1 + π/2. Setting

t1 = 1 + π/2 in y(t) and letting γ → 0, we find that y1 → 1. These conclusions agree with
the case γ = 0, for which it is straightforward to show that y(t) = sin(t − 1)u1 (t).
15.(a) The solution of this initial value problem is
√ !
10k 3 11
y(t) = √ e−(t−1)/10 sin (t − 1) u1 (t).
3 11 10

This function is a multiple of the answer in problem 14(d), with γ = 1/5. Therefore, the
peak value occurs at t1 ≈ 2.4780. The maximum value is calculated as y(2.4780) ≈ 0.8626k.
We find that the appropriate value of k is k1 = 6/0.8626 ≈ 6.9557.
(b) Based on problem 14(d) with γ = 1/10, the solution is
√ !
20k −(t−1)/20 399
y(t) = √ e sin (t − 1) u1 (t).
399 20

Since this function is a multiple of the solution in problem 14(c), we have t1 ≈ 2.5227 with
y(t1 ) ≈ 0.9267k. The solution reaches a maximum value of y = 6 for k1 = 6/0.9267 ≈ 6.4746.
(c) Similar to problem 14(d), for 0 < γ < 1, the solution is given by

2k p 
y(t) = p e−γ(t−1)/2 sin 1 − γ 2 /4(t − 1) u1 (t).
4 − γ2

As γ → 0, we see that y 0 (t) = 0 at t1 → 1 + π/2. Setting t1 = 1 + π/2 in y(t) and

letting γ → 0, we find that y1 → k. Requiring that the peak value remains at y = 6, the
limiting value of k is k1 = 6. These conclusions agree with the case γ = 0 for which it is
straightforward to show that y(t) = k sin(t − 1)u1 (t).
16.(a) Applying the Laplace transform and using the initial conditions, this equation becomes

1 e−(3−k)s e−(3+k)s
 
2
[s + 1]Y (s) = − .
2k s s

Therefore,

1 e−(3−k)s − e−(3+k)s
   
1 −(3−k)s −(3+k)s 1 s
Y (s) = = (e −e ) − .
2k s(s2 + 1) 2k s s2 + 1

Thus we obtain that

1
y(t) = φ(t, k) = [u3−k (t)[1 − cos(t − 3 + k)] − u3+k (t)[1 − cos(t − 3 − k)]] .
2k
368 CHAPTER 5. THE LAPLACE TRANSFORM

(b) For any t < 3, φ(t, k) = 0 as long as 3 − k > t. If t ≥ 3, then for 3 + k < t,
1
φ(t, k) = − [cos(t − 3 + k) − cos(t − 3 − k)].
2k
It follows that
cos(t − 3 + k) − cos(t − 3 − k)
lim φ(t, k) = lim − = sin(t − 3).
k→0 k→0 2k
Therefore,
lim φ(t, k) = sin(t − 3)u3 (t).
k→0

(c) Replacing the nonhomogeneous term with δ(t − 3), the equation becomes
y 00 + y = δ(t − 3).
Applying the Laplace transform and using the initial conditions, we have
[s2 + 1]Y (s) = e−3s .
Therefore,
e−3s
Y (s) = .
s2 + 1
It follows that the solution is φ0 (t) = sin(t − 3)u3 (t).
(d)

17.(b) Applying the Laplace transform to the equation and using the initial conditions, we
have
20
X
2
[s + 1]Y (s) = e−kπs .
k=1
Therefore,
20
1 X −kπs
Y (s) = 2 e ,
s + 1 k=1
which implies
20
X
y(t) = sin(t − kπ)ukπ (t).
k=1
5.7. IMPULSE FUNCTIONS 369

(c) After the sequence of impulses ends, the oscillator returns to equilibrium.
18.(b) Applying the Laplace transform to the equation and using the initial conditions, we
have
20
X
2
[s + 1]Y (s) = (−1)k+1 e−kπs .
k=1
Therefore,
20
1 X
Y (s) = 2 (−1)k+1 e−kπs ,
s + 1 k=1
which implies
20
X
y(t) = (−1)k+1 sin(t − kπ)ukπ (t).
k=1

(c) After the sequence of impulses ends, the oscillator continues to oscillate at a constant
amplitude.
19.(b) Applying the Laplace transform to the equation and using the initial conditions, we
have
20
X
2
[s + 1]Y (s) = e−(kπ/2)s .
k=1
Therefore,
20
1 X −(kπ/2)s
Y (s) = 2 e ,
s + 1 k=1
370 CHAPTER 5. THE LAPLACE TRANSFORM

which implies
20
X
y(t) = sin(t − kπ/2)ukπ/2 (t).
k=1

(c) After the sequence of impulses ends, the oscillator returns to equilibrium.
20.(b) Applying the Laplace transform to the equation and using the initial conditions, we
have
20
X
2
[s + 1]Y (s) = (−1)k+1 e−(kπ/2)s .
k=1

Therefore,
20
1 X
Y (s) = 2 (−1)k+1 e−(kπ/2)s ,
s + 1 k=1
which implies
20
X
y(t) = (−1)k+1 sin(t − kπ/2)ukπ/2 (t).
k=1

(c) After the sequence of impulses ends, the oscillator returns to equilibrium.
21.(b) Applying the Laplace transform to the equation and using the initial conditions, we
have
X20
2
[s + 1]Y (s) = e−(2k−1)πs .
k=1
5.7. IMPULSE FUNCTIONS 371

Therefore,
20
1 X −(2k−1)πs
Y (s) = 2 e ,
s + 1 k=1

which implies
20
X
y(t) = sin(t − (2k − 1)π)u(2k−1)π (t).
k=1

(c) After the sequence of impulses ends, the oscillator continues to oscillate at a constant
amplitude.
22.(b) Applying the Laplace transform to the equation and using the initial conditions, we
have
20
X
2
[s + 1]Y (s) = (−1)k+1 e−(15k/4)s .
k=1

Therefore,
20
1 X
Y (s) = 2 (−1)k+1 e−(15k/4)s ,
s + 1 k=1

which implies
20
X
y(t) = (−1)k+1 sin(t − 15k/4)u15k/4 (t).
k=1
372 CHAPTER 5. THE LAPLACE TRANSFORM

(c) After the sequence of impulses ends, the oscillator continues to oscillate at a constant
amplitude.
23.(b) Applying the Laplace transform to the equation and using the initial conditions, we
have
20
X
2
[s + 0.1s + 1]Y (s) = (−1)k+1 e−kπs .
k=1

Therefore,
20 20
1 X 1 X
Y (s) = 2 (−1)k+1 e−kπs = (−1)k+1 e−kπs ,
s + 0.1s + 1 k=1 (s + 1/20)2 + (399/400) k=1

which implies
20 √ !
X 20 −(t−kπ)/20 399
y(t) = (−1)k+1 √ e sin (t − kπ) ukπ (t).
k=1
399 20

(c) After the sequence of impulses ends, the solution oscillates to zero due to the damping
term.
24.(b) Applying the Laplace transform to the equation and using the initial conditions, we
have
15
X
[s2 + 0.1s + 1]Y (s) = e−(2k−1)πs .
k=1

Therefore,
15 15
1 X
−(2k−1)πs 1 X
Y (s) = e = e−(2k−1)πs ,
s2 + 0.1s + 1 k=1 (s + 1/20)2 + (399/400) k=1

which implies
15 √ !
X 20 −(t−(2k−1)π)/20 399
y(t) = √ e sin (t − (2k − 1)π) u(2k−1)π (t).
k=1
399 20
5.7. IMPULSE FUNCTIONS 373

(c) After the sequence of impulses ends, the solution oscillates to zero due to the damping
term.
25.(a) A fundamental set of solutions to the homogeneous problem is y1 (t) = e−t cos t and
y2 (t) = e−t sin t. Based on problem 22 from Section 4.8, a particular solution is given by

t
y1 (s)y2 (t) − y1 (t)y2 (s)
Z
yp (t) = f (s) ds.
0 W (y1 , y2 )(s)

Using the functions y1 , y2 above, we see that

t t
e−(s−t) [cos(s) sin(t) − sin(s) cos(t)]
Z Z
yp (t) = f (s) ds = e−(t−s) sin(t − s)f (s) ds.
0 e−2s 0

Given the fact that the initial conditions are zero, the solution of this initial value problem
is
Z t
yp (t) = e−(t−s) sin(t − s)f (s) ds.
0

(b) If f (t) = δ(t − π), then for t < π, y(t) = 0. On the other hand, for t > π,
Z t
y(t) = e−(t−s) sin(t − s)δ(s − π) ds = e−(t−π) sin(t − π).
0

Therefore, y(t) = uπ (t)e−(t−π) sin(t − π).

(c) Applying the Laplace transform to the ODE with f (t) = δ(t − π), we conclude that the
Laplace transform satisfies

e−πs e−πs
Y (s) = = .
s2 + 2s + 2 (s + 1)2 + 1

Therefore,
y(t) = uπ (t) sin(t − π).
374 CHAPTER 5. THE LAPLACE TRANSFORM

5.8 Convolution Integrals and Their Applications

1.(a) We compute:
Z t Z t Z t
f ∗ (g1 + g2 )(t) = f (t − τ )[g1 + g2 ](τ ) dτ = f (t − τ )g1 (τ ) dτ + f (t − τ )g2 (τ ) dτ
0 0 0
= f ∗ g1 (t) + f ∗ g2 (t).

(b) We compute:
Z t Z t Z τ 
f ∗ (g ∗ h)(t) = f (t − τ )[g ∗ h(τ )] dτ = f (t − τ ) g(τ − η)h(η) dη dτ
0 0 0
Z tZ τ
= f (t − τ )g(τ − η)h(η) dη dτ.
0 0

Interchanging the order of integration, we have

Z tZ τ Z tZ t
f (t − τ )g(τ − η)h(η) dη dτ = f (t − τ )g(τ − η)h(η) dτ dη
0 0 0 η
Z t Z t 
= f (t − τ )g(τ − η) dτ h(η) dη.
0 η

Now letting τ − η = u, we obtain

Z t Z t−η
f (t − τ )g(τ − η) dτ = f (t − η − u)g(u) du = f ∗ g(t − η).
η 0

Therefore, Z t Z t
f (t − τ )[g ∗ h(τ )] dτ = [f ∗ g(t − τ )]h(τ ) dτ.
0 0

2. For f (t) = sin t,

Z t
1 t
Z
1
f ∗ f (t) = sin(t − τ ) sin τ dτ = [cos(t − 2τ ) − cos t] dτ = (sin t − t cos t).
0 2 0 2
Clearly, f ∗ f is not necessarily nonnegative.
3. L[t4 ] = 24/s5 and L[cos 6t] = s/(s2 + 36). Therefore,
Z t     
4 24 s 24
L (t − τ ) cos 6τ dτ = 5
· 2
= 4 2 .
0 s s + 36 s (s + 36)

4. L[e−t ] = 1/(s + 1) and L[sin t] = 1/(s2 + 1). Therefore,

Z t     
−(t−τ ) 1 1 1
L e sin τ dτ = · 2
= .
0 s+1 s +1 (s + 1)(s2 + 1)
5.8. CONVOLUTION INTEGRALS AND THEIR APPLICATIONS 375

5. L[t] = 1/s2 and L[et ] = 1/(s − 1). Therefore,

Z t     
τ 1 1 1
L (t − τ )e dτ = 2
· = 2 .
0 s s−1 s (s − 1)

6. L[sin t] = 1/(s2 + 1) and L[cos t] = s/(s2 + 1). Therefore,

Z t     
1 s s
L sin(t − τ ) cos τ dτ = · = .
0 s2 + 1 s2 + 1 (s2 + 1)2

7. L−1 [1/s3 ] = t2 /2 and L−1 [1/(s2 + 1)] = sin t. Therefore,

  Z t
−1 1 1
L 3 2
= (t − τ )2 sin τ dτ.
s (s + 1) 0 2

8. L−1 [1/(s + 1)] = e−t and L−1 [2/(s2 + 4)] = cos 2t. Therefore,
  Z t
−1 s
L = e−(t−τ ) cos 2τ dτ.
(s + 1)(s2 + 4) 0

1 1
9. L−1 [1/(s + 3)4 ] = t3 e−3t and L−1 [1/(s2 + 4)] = sin 2t. Therefore,
6 2
  Z t
−1 1 1
L 4 2
= (t − τ )3 e−3(t−τ ) sin 2τ dτ.
(s + 3) (s + 4) 0 12

10. Let g(t) = L−1 [G(s)]. Since L−1 [1/(s2 + 1)] = sin t,
  Z t
−1 G(s)
L = g(t − τ ) sin τ dτ.
s2 + 1 0

11. Consider
1 1 1
F (s) = = · = G(s) · G(s);
(s2 + 1)2 s2 + 1 s2 + 1
then Z t
−1
L (F (s)) = g(t − τ )g(τ ) dτ,
0

where g(t) = L−1 (G(s)) = sin t. Therefore,

Z t
−1
L (F (s)) = sin(t − τ ) sin τ dτ.
0

12. Let
5s
F (s) = = G(s) · H(s),
(s2 + 25)2
376 CHAPTER 5. THE LAPLACE TRANSFORM

where G(s) = s/(s2 + 25) and H(s) = 5/(s2 + 25). Therefore,

Z t
−1
L (F (s)) = g(t − τ )h(τ ) dτ
0

where g(t) = L−1 (G(s)) = cos 5t and h(t) = L−1 (H(s)) = sin 5t. Therefore,
Z t
−1
L (F (s)) = cos 5(t − τ ) sin 5τ dτ.
0

13.(a) We know that Z t

f ∗g = (t − τ )m τ n dτ.
0
Introducing a new variable u such that τ = t − tu, we have dτ = −t du. Therefore,
Z t Z 0 Z 1
m n m n m+n+1
(t − τ ) τ dτ = − (tu) (t − tu) t du = t um (1 − u)n du.
0 1 0

(b) From part (a),

1
(m + n + 1)! 1 m
 Z  Z
m+n+1 m n
L[f ∗ g] = L t u (1 − u) du = u (1 − u)n du.
0 sm+n+2 0

By the Convolution Theorem, we know that L[f ∗ g] = L[f ] · L[g]. Therefore, calculating
L[f ] · L[g], we have
m! n!
L[f ] · L[g] = m+1 · n+1 .
s s
Equating L[f ∗ g] and L[f ] · L[g], we obtain
(m + n + 1)! 1 m
Z
m! n!
m+n+2
u (1 − u)n du = m+1 · n+1 .
s 0 s s
Therefore, we conclude that
Z 1
m!n!
um (1 − u)n du = .
0 (m + n + 1)!

(c) For the case when m and n are positive numbers, but not necessarily integers, we have
Z 1
Γ(m + n + 2) 1 m
  Z
m+n+1 m n
L[f ∗ g] = L t u (1 − u) du = m+n+2
u (1 − u)n du.
0 s 0

Further,
Γ(m + 1) Γ(n + 1)
L[f ] · L[g] = · .
sm+1 sn+1
Equating L[f ∗ g] and L[f ] · L[g], we obtain
Γ(m + n + 2) 1 m
Z
Γ(m + 1) Γ(n + 1)
m+n+2
u (1 − u)n du = · .
s 0 sm+1 sn+1
5.8. CONVOLUTION INTEGRALS AND THEIR APPLICATIONS 377

Therefore, we conclude that

Z 1
Γ(m + 1)Γ(n + 1)
um (1 − u)n du = .
0 Γ(m + n + 2)

14. Applying the Laplace transform to the equation, we have

[s2 Y (s) − sy(0) − y 0 (0)] + ω 2 Y (s) = L[g(t)].

Letting G(s) = L[g(t)] and applying the initial conditions, we get

[s2 + ω 2 ]Y (s) = 1 + G(s).

Therefore,
1 G(s)
Y (s) = + 2 .
+ω s2
2 s + ω2
Now   Z
−1 G(s) 1
L = sin ω(t − τ )g(τ ) dτ.
s2 + ω 2 ω
Thus Z t
1 1
y(t) = sin ωt + sin ω(t − τ )g(τ ) dτ.
ω ω 0

15. Applying the Laplace transform to the equation, we have

α
[s2 Y (s) − sy(0) − y 0 (0)] + 6[sY (s) − y(0)] + 25Y (s) = .
s2 + α 2
Applying the initial conditions, we get
α
[s2 + 6s + 25]Y (s) = .
s2 + α 2
Therefore,
1 α 1 α
Y (s) = · 2 = · 2 .
s2 + 6s + 25 s + α 2 (s + 3) + 16 s + α2
2

Thus Z t
1
y(t) = e−3(t−τ ) sin 4(t − τ ) sin ατ dτ.
4 0

16. Applying the Laplace transform to the equation, we have

4[s2 Y (s) − sy(0) − y 0 (0)] + 4[sY (s) − y(0)] + 17Y (s) = G(s),

where G(s) = L[g(t)]. Applying the initial conditions, we get

[4s2 + 4s + 17]Y (s) = G(s).

Therefore,
G(s) 1 G(s)
Y (s) = = .
4s2 + 4s + 17 4 (s + 1/2)2 + 4
378 CHAPTER 5. THE LAPLACE TRANSFORM

Thus Z t
1
y(t) = e−(t−τ )/2 sin(2(t − τ ))g(τ ) dτ.
8 0

17. Applying the Laplace transform to the equation, we have

5 1 e−πs
[s2 Y (s) − sy(0) − y 0 (0)] + [sY (s) − y(0)] + Y (s) = − .
4 s s
Applying the initial conditions, we get

2 1 e−πs
[s + s + 5/4]Y (s) = s + − .
s s
Therefore,

s 1 1 − e−πs
Y (s) = + ·
s2 + s + 5/4 s2 + s + 5/4 s
s + 1/2 1/2 1 1 − e−πs
= − + ·
(s + 1/2)2 + 1 (s + 1/2)2 + 1 (s + 1/2)2 + 1 s

Thus Z t
−t/2 1
y(t) = e cos(t) − e−t/2 sin(t) + e−(t−τ )/2 sin(t − τ )(1 − uπ (τ )) dτ.
2 0

18. Applying the Laplace transform to the equation, we have

[s2 Y (s) − sy(0) − y 0 (0)] + 4[sY (s) − y(0)] + 4Y (s) = G(s),

where G(s) = L[g(t)]. Applying the initial conditions, we get

[s2 + 4s + 4]Y (s) = 2s + 5 + G(s).

Therefore,
2s + 5 G(s)
Y (s) = 2
+ .
(s + 2) (s + 2)2
Using partial fractions, we obtain
2s + 5 2 1
2
= + ,
(s + 2) s + 2 (s + 2)2

so we can conclude that

Z t
−2t −2t
y(t) = 2e + te + (t − τ )e−2(t−τ ) g(τ ) dτ.
0

19. Applying the Laplace transform to the equation, we have

s
[s2 Y (s) − sy(0) − y 0 (0)] + 3[sY (s) − y(0)] + 2Y (s) = .
s2 + α2
5.8. CONVOLUTION INTEGRALS AND THEIR APPLICATIONS 379

Applying the initial conditions, we get

s
[s2 + 3s + 2]Y (s) = s + 3 + .
s2 + α2
Therefore,
s+3 s
Y (s) = + 2 .
s2 + 3s + 2 (s + 3s + 2)(s2 + α2 )
Using partial fractions, we obtain
s+3 2 1
= −
s2 + 3s + 2 s+1 s+2
and
1 1 1
= − .
s2 + 3s + 2 s+1 s+2
Thus we conclude that
Z t
−t −2t
y(t) = 2e −e + (e−(t−τ ) − e−2(t−τ ) ) cos ατ dτ.
0

20. Applying the Laplace transform to the equation and using the initial conditions, we have

[s4 − 16]Y (s) = G(s).

Therefore,
G(s)
Y (s) = .
s4 − 16
Using partial fractions, we can write
 
1 1 1 1
= − .
s4 − 16 8 s2 − 4 s2 + 4

Thus we conclude that

Z t
1
y(t) = [sinh 2(t − τ ) − sin 2(t − τ )]g(τ ) dτ.
16 0

21. Applying the Laplace transform to the equation, we have

[s4 Y (s) − s3 y(0) − s2 y 0 (0) − sy 00 (0) − y 000 (0)] + 17[s2 Y (s) − sy(0) − y 0 (0)] + 16Y (s) = G(s).

Applying the initial conditions, we get

[s4 + 17s2 + 16]Y (s) = 2s3 + 34s + G(s).

Therefore,
2s3 + 34s G(s)
Y (s) = + 4 .
s + 17s + 16 s + 17s2 + 16
4 2
380 CHAPTER 5. THE LAPLACE TRANSFORM

Using partial fractions, we write

2s3 + 34s 32s 2s
= −
s4 + 17s2 + 16 15(s2 + 1) 15(s2 + 16)
and
1 1 1
= − .
s4 + 17s2 + 16 15(s2 + 1) 15(s2 + 16)
Therefore, we can conclude that
Z t
32 2 1
y(t) = cos t − cos 4t + [4 sin(t − τ ) − sin 4(t − τ )]g(τ ) dτ.
15 15 60 0

22.(a) The identity Z t 

0 t<0
δ(τ ) dτ =
−∞ 1 t>0
implies that Z t
δ(τ ) dτ = u(t),
−∞

thus δ(t) = u0 (t).

(b) Applying the Laplace transform and the initial conditions, we have
1
[s2 + 2δ + 1]Y (s) = .
s
Therefore,
1 1
Y (s) = · .
(s + δ)2 2
+ (1 − δ ) s
Thus we conclude that
1
Z t √
y(t) = √ e−δτ sin( 1 − δ 2 τ ) dτ
1 − δ2 0

for 0 < δ < 1. For δ = 1, it is straightforward to obtain the solution

y(t) = −te−t − e−t + 1.

5.8. CONVOLUTION INTEGRALS AND THEIR APPLICATIONS 381

23. Taking the Laplace transform of the integral equation, we have

Z t 
L[φ(t)] + L k(t − ξ)φ(ξ) dξ = L[f (t)],
0

which implies
L[φ(t)] + L[k(t)] · L[φ(t)] = L[f (t)].
Therefore, L[φ(t)](1 + L[k(t)]) = L[f (t)], and then
L[f (t)] F (s)
Φ(s) = L[φ(t)] = = .
1 + L[k(t)] 1 + K(s)

24.(a) Applying the result in problem 23 above, we have

L[sin 2t] 2/(s2 + 4) 2s2
L[φ(t)] = = = .
1 + L[t] 1 + 1/s2 (s2 + 1)(s2 + 4)
Using partial fractions, we can write
2s2
 
2 4 1
= − .
(s2 + 1)(s2 + 4) 3 s2 + 4 s2 + 1
Therefore,
2
φ(t) = (2 sin 2t − sin t).
3
(b) Differentiating the equation once, we get
Z t
0
φ (t) + φ(ξ) dξ = 2 cos 2t.
0

Differentiating the equation again, we have

φ00 (t) + φ(t) = −4 sin 2t.

Using equation (i), we see that φ(0) = sin(2 · 0) = 0. Further, using the equation for the first
derivative of (i), we see that φ0 (0) = 2 cos(2 · 0) = 2.
(c) Letting Φ(s) = L[φ(t)] and taking the Laplace transform of the equation in part (b), we
get
8
[s2 Φ(s) − sφ(0) − φ0 (0)] + Φ(s) = − 2 .
s +4
Applying the initial conditions, we have
8
[s2 + 1]Φ(s) = 2 − .
s2 +4
Therefore,
 
2 8 2 8 1 1 4 2 2 1
Φ(s) = 2 − 2 2
= 2 + 2
− 2 = 2
− 2 ,
s + 1 (s + 1)(s + 4) s +1 3 s +4 s +1 3s +4 3s +1
382 CHAPTER 5. THE LAPLACE TRANSFORM

which implies that

4 2
φ(t) = sin 2t − sin t.
3 3
25.(a) Taking the Laplace transform of the equation, we have
1 1
Φ(s) + 2
Φ(s) = .
s s
Therefore,
s
Φ(s) = ,
s2 +1
which implies that φ(t) = cos t.
(b) Differentiating the equation once, we have
Z t
0
φ (t) + φ(ξ) dξ = 0.
0

Differentiating the equation again, we have

φ00 (t) + φ(t) = 0.

Using equation (i), we see that φ(0) = 1. Further, using the equation for the first derivative
of (i), we see that φ0 (0) = 0.
(c) Letting Φ(s) = L[φ(t)] and taking the Laplace transform of the equation in part (ii), we
have
[s2 Φ(s) − sφ(0) − φ0 (0)] + Φ(s) = 0.
Applying the initial conditions, we get [s2 + 1]Φ(s) = s. Therefore,
s
Φ(s) = ,
s2 +1
which implies that φ(t) = cos t.
26.(a) Taking the Laplace transform of the equation, we have
1 1
Φ(s) − 2
Φ(s) = .
s s
Therefore,
s
Φ(s) = ,
s2 −1
which implies φ(t) = cosh t.
(b) Differentiating the equation once, we have
Z t
0
φ (t) − φ(ξ) dξ = 0.
0

Differentiating the equation again, we have

φ00 (t) − φ(t) = 0.

5.8. CONVOLUTION INTEGRALS AND THEIR APPLICATIONS 383

Using equation (i), we see that φ(0) = 1. Further, using the equation for the first derivative
of (i), we see that φ0 (0) = 0.
(c) Letting Φ(s) = L[φ(t)] and taking the Laplace transform of the equation in part (b), we
have
[s2 Φ(s) − sφ(0) − φ0 (0)] − Φ(s) = 0.
Applying the initial conditions, we have [s2 − 1]Φ(s) = s. Therefore,
s
Φ(s) = ,
s2 −1

which implies φ(t) = cosh t.

27.(a) Taking the Laplace transform of the equation, we have

2s 1
Φ(s) + Φ(s) = .
s2 +1 s+1
Therefore,
s2 + 1 1 2 2
Φ(s) = 3
= − 2
+ ,
(s + 1) s + 1 (s + 1) (s + 1)3
which implies that φ(t) = e−t − 2te−t + t2 e−t .
(b) Differentiating the equation once, we have
Z t
0
φ (t) + 2φ(t) − 2 sin(t − ξ)φ(ξ) dξ = −e−t .
0

Differentiating the equation again, we have

Z t
00 0
φ (t) + 2φ (t) − 2 cos(t − ξ)φ(ξ) dξ = e−t .
0

Using the original equation, this gives us

φ00 (t) + 2φ0 (t) + φ(t) = 2e−t .

Using equation (i), we see that φ(0) = 1. Further, using the equation for the first derivative
of (i), we see that φ0 (0) = −3.
(c) Letting Φ(s) = L[φ(t)] and taking the Laplace transform of the equation in part (ii), we
have
s 1
[s2 Φ(s) − sφ(0) − φ0 (0)] + 2[sΦ(s) − φ(0)] − 2 2 Φ(s) = .
s +1 s+1
Applying the initial conditions, we get
 
2 2s 1
s + 2s − 2 Φ(s) = s − 1 + .
s +1 s+1
384 CHAPTER 5. THE LAPLACE TRANSFORM

Therefore,
(s − 1)(s2 + 1) s2 + 1
Φ(s) = +
s4 + 2s3 + s2 (s + 1)(s4 + 2s3 + s2 )
1 4 3 2 1 2 3 3 2
= − 2− 2
+ − + 2+ 2
− + +
s (s + 1) s s+1 s (s + 1) s s + 1 (s + 1)3
2 1 2
= − 2
+ + ,
(s + 1) s + 1 (s + 1)3
which implies φ(t) = −2te−t + e−t + t2 e−t .
28.(a) Taking the Laplace transform of the equation, we have
1 1
sΦ(s) + 2
Φ(s) = 2 .
s s
Therefore,
 
1 2−s 1 1 1 s − 1/2 1/2 1 1
Φ(s) = 3 = 2
+ =− 2
+ 2
+ ,
s +1 3(s − s + 1) 3 s + 1 3 (s − 1/2) + 3/4 (s − 1/2) + 3/4 3 s + 1
which implies
√ ! √ !
1 3t 1 3t 1
φ(t) = − et/2 cos + √ et/2 sin + e−t .
3 2 3 2 3

(b) Differentiating the equation once, we have

Z t
00
φ (t) + φ(ξ) dξ = 1.
0

Differentiating the equation again, we have

φ000 (t) + φ(t) = 0.

Using the equation for φ0 , we see that φ0 (0) = 0. Further, using the equation for φ00 , we see
that φ00 (0) = 1.
(c) Letting Φ(s) = L[φ(t)] and taking the Laplace transform of the equation in part (b), we
have
[s3 Φ(s) − s2 φ(0) − sφ0 (0) − φ00 (0)] + Φ(s) = 0.
Applying the initial conditions, we have [s3 + 1]Φ(s) = 1. Therefore,
1
Φ(s) = .
s3 +1
Using the same analysis as in part (a), we see that
√ ! √ !
1 t/2 3t 1 3t 1
φ(t) = − e cos + √ et/2 sin + e−t .
3 2 3 2 3
5.8. CONVOLUTION INTEGRALS AND THEIR APPLICATIONS 385

29.(a) Taking the Laplace transform of the equation, we have

1 1
sΦ(s) − 1 − Φ(s) = − .
s3 s2
Therefore,
s
Φ(s) = ,
s2 +1
which implies φ(t) = cos t.
(b) Differentiating the equation once, we have
Z t
00
φ (t) − (t − ξ)φ(ξ) dξ = −1.
0

Now Z t Z t Z t
− (t − ξ)φ(ξ) dξ = −t φ(ξ) dξ + ξφ(ξ) dξ;
0 0 0
using the product rule for differentiation and the fundamental theorem of calculus, after
differentiating the equation again, we get
Z t
000
φ (t) − φ(ξ) dξ = 0.
0

Differentiating again, we see that

φ0000 (t) − φ(t) = 0.

Using the equations above for the derivatives of φ, we see that φ(0) = 1, φ0 (0) = 0, φ00 (0) =
−1 and φ000 (0) = 0.
(c) Letting Φ(s) = L[φ(t)] and taking the Laplace transform of the equation in part (ii), we
have
[s4 Φ(s) − s3 φ(0) − s2 φ0 (0) − sφ00 (0) − φ000 (0)] − Φ(s) = 0.
Applying the initial conditions, we get [s4 − 1]Φ(s) = s3 − s. Therefore,

s3 − s s(s2 − 1) s
Φ(s) = 4
= 2 2
= 2 .
s −1 (s − 1)(s + 1) s +1

Therefore, φ(t) = cos t.

30.(a) Taking the Laplace transform of the equation, we have
1
sΦ(s) − 1 + Φ(s) = Φ(s).
s2 +1
Therefore,
√
s2 + 1 1 1 1 2 3/2
Φ(s) = 2
= − 2 = −√ ,
s(s + s + 1) s s +s+1 s 3 (s + 1/2)2 + 3/4
386 CHAPTER 5. THE LAPLACE TRANSFORM

which implies
√ !
2 3
φ(t) = 1 − √ e−t/2 sin t .
3 2

(b) Differentiating the equation once, we have

Z t
00 0
φ (t) + φ (t) = cos(t − ξ)φ(ξ) dξ.
0

Now Z t Z t Z t
cos(t − ξ)φ(ξ) dξ = cos t cos ξ φ(ξ) dξ + sin t sin ξ φ(ξ) dξ;
0 0 0
using the product rule for differentiation, the fundamental theorem of calculus, and the fact
that sin t cos ξ − cos t sin ξ = sin(t − ξ), after differentiating the equation again, we get
Z t
000 00
φ (t) + φ (t) = φ(t) − sin(t − ξ)φ(ξ) dξ = −φ0 (t).
0

Therefore,
φ000 (t) + φ00 (t) + φ0 (t) = 0.
Using the equations above for the derivatives of φ, we see that φ0 (0) = −1 and φ00 (0) = 1.
(c) Letting Φ(s) = L[φ(t)] and taking the Laplace transform of the equation in part (b), we
have

[s3 Φ(s) − s2 φ(0) − sφ0 (0) − φ00 (0)] + [s2 Φ(s) − sφ(0) − φ0 (0)] + [sΦ(s) − φ(0)] = 0.

Applying the initial conditions, we get [s3 + s2 + s]Φ(s) = s2 + 1. Therefore,

s2 + 1
Φ(s) = .
s3 + s2 + s
Using the analysis from part (a), we conclude that
√ !
2 3
φ(t) = 1 − √ e−t/2 sin t .
3 2

31.(a) We note that

Z b  
f (y) 1
√ dy = √ ∗f (b).
0 b−y y
Taking the Laplace transform of both sides of this equation, we see that
 
T0 1 1
= √ F (s) · L √ .
s 2g y
Then using the fact that   r
1 Γ(1/2) π
L √ = 1/2 = ,
y s s
5.8. CONVOLUTION INTEGRALS AND THEIR APPLICATIONS 387

we conclude that
T0 1 √
= √ F (s) · πs,
s 2g
which implies
r
2g T0
F (s) = ·√ .
π s
Taking the inverse transform, we see that
r
T0 2g
f (y) = .
π y

(b) Combining equations (i) and (iv), we see that

2
2gT02

dx
=1+ .
π2y dy

Solving this equation for dx/dy, we see that

r
dx 2α − y
= ,
dy y

where α = gT02 /π 2 .
(c) Consider the change of variables y = 2α sin2 (θ/2). Using the chain rule, we have

dy dθ
= 2α sin(θ/2) cos(θ/2) ·
dx dx

and
dx 1 dx
= · .
dy 2α sin(θ/2) cos(θ/2) dθ
It follows that
s
dx cos2 (θ/2)
= 2α sin(θ/2) cos(θ/2) = 2α cos2 (θ/2) = α + α cos(θ).
dθ sin2 (θ/2)

Now integrating this equation for dx/dθ, we have

x(θ) = αθ + α sin(θ) + C.

Since the curve passes through the origin, we have x(0) = 0 = y(0). Therefore, C = 0, and
x(θ) = αθ + α sin(θ). We also have y(θ) = 2α sin2 (θ/2) = α − α cos(θ).
388 CHAPTER 5. THE LAPLACE TRANSFORM

5.9 Linear Systems and Feedback Control

1. Let Y1 represent the output before the junction point. Let Y2 represent the output after
the junction point. Let U1 represent the output of controller G1 and let U2 represent the
output of controller G2 . Then we have the following system of equations
Y1 = H1 [F − U1 ]
U1 = G1 Y1
Y2 = Y1 + U2
U2 = G2 [F − U1 ].
Solving the equation for Y1 , we get Y1 = H1 F − H1 G1 Y1 , thus
H1 F
Y1 = .
1 + H1 G1
Then Y2 = Y1 + U2 implies that
 
H1 F G1 H1 F H1 F + G2 F
Y2 = + G2 F − = .
1 + H1 G1 1 + H1 G1 1 + H1 G1

2. Let Y1 represent the output after H1 , but before the junction point. Let Y2 represent
the output between H2 and H3 . Let Y3 represent the output after H3 . Let U1 represent the
output of controller G1 and let U2 represent the output of controller G2 . Then we have the
following system of equations
Y1 = H1 [F − U2 ]
Y2 = H2 [Y1 − U1 ]
Y3 = H3 Y2
U1 = G1 Y3
U2 = G2 Y2 .
Replacing U1 and U2 in the first two equations, we have
Y1 = H1 [F − G2 Y2 ]
Y2 = H2 [Y1 − G1 Y3 ].
Using the fact that Y3 = H3 Y2 , we can eliminate Y2 in the above equations. We conclude
that
 
Y3
Y1 = H1 F − G2 H1
H3
Y3
= H2 Y1 − G1 H2 Y3 .
H3
Substituting Y1 into the second of these equations, we conclude that
 
Y3 H1 H3 F − G2 H1 Y3
= H2 − G1 H2 Y3 .
H3 H3
5.9. LINEAR SYSTEMS AND FEEDBACK CONTROL 389

Solving this equation for Y3 , we have

H1 H2 H3 F
Y3 = .
1 + G2 H1 H2 + G1 H2 H3
Therefore, we conclude that the overall transfer function is
Y3 H1 H2 H3
H(s) = = .
F 1 + G2 H1 H2 + G1 H2 H3

3. Consider the function h(t) = tk eαt . We will prove that

Z ∞
k!
|h(t)| dt ≤
0 |α|k+1
by induction on k. First, for k = 1, we have
Z ∞ b Z ∞ α b
αt teαt  e t eαt  1
te dt = lim − dt = − lim 2  = 2 .
0 b→∞ α 
0 α b→∞ α α
0 0

Now assume that Z ∞

k!
tk eαt dt ≤
0 |α|k+1
to prove that Z ∞
(k + 1)!
tk+1 eαt dt ≤ .
0 |α|k+2
Integrating by parts, we see that
∞
b Z ∞ αt Z ∞ αt
tk+1 eαt  e (k + 1)tk
Z
k+1 αt e k
t e dt = lim − (k + 1)t dt = − dt.
0 b→∞ α 0 0 α 0 α

From our inductive assumption,

Z ∞ αt k

 e (k + 1)t  k+1 k! (k + 1)!
 dt ≤ · k+1
= .

0 α α |α| |α|k+2

Further, |tk eαt cos(βt)| ≤ tk eαt and |tk eαt sin(βt)| ≤ tk eαt . Therefore, the same bounds hold
on integrals involving those functions.
4. We will prove this statement by induction. Consider the case when pn (x) is a first-degree
polynomial. That is, p1 (x) = x + a. In this case, there is exactly one real root, x = −a. If
this root is negative, then −a < 0 which implies that a > 0. Now we will proceed inductively.
Let
pn (x) = xn + an−1 xn−1 + . . . + a1 x + a0 .
Let c1 , . . . , cn be the n roots of pn . Our inductive assumption is that if Re(ci ) < 0 for
i = 1, . . . , n, then a0 , . . . , an−1 are all positive. We now consider a polynomial of degree
n + 1. Consider
pn+1 (x) = xn+1 + an xn + . . . + a1 x + a0 .
390 CHAPTER 5. THE LAPLACE TRANSFORM

Let c1 , . . . , cn+1 denote the n roots of pn . We assume that Re(ci ) < 0 for i = 1, . . . , n + 1.
We will show that a0 , . . . , an > 0. Since c1 is one of the roots, then we can write

pn+1 (x) = (x − c1 )(xn + bn−1 xn−1 + . . . + b1 x + b0 ).

Let
qn (x) ≡ xn + bn−1 xn−1 + . . . + b1 x + b0 .
Now qn is an nth degree polynomial whose roots are given by c2 , . . . , cn+1 . Since all of
these roots satisfy Re(ci ) < 0, by our inductive assumption, b0 , . . . , bn−1 > 0. Therefore,
multiplying (x − c1 ) by qn , we have

pn+1 (x) = (x − c1 )(xn + bn−1 xn−1 + . . . + b1 x + b0 )

= xn+1 + (bn−1 − c1 )xn + (bn−2 − bn−1 c1 )xn−1 + . . . + (b0 − b1 c1 )x − b0 c1 .

Under the assumption that b0 , . . . , bn−1 > 0 and Re(c1 ) < 0, we conclude that each of these
coefficients is positive.
5. The associated Routh table is given by

s3 1 a1 0
s2 a2 a0 0
a1 a2 − a0
s 0
a2
s0 a0

In order for all the roots to have negative real parts, we need all entries in the first column to
be positive. Therefore, we need (1) a2 > 0, (2) (a1 a2 − a0 )/a2 > 0 and (3) a0 > 0. Therefore,
all the roots will lie in the left half plane if and only if (1) a2 > 0, (2) a1 a2 > a0 and (3)
a0 > 0.
6. The associated Routh table is given by

s3 1 2 0
s2 5 13 0
s −3/5 0
s0 13

There are two sign changes in the first column. Therefore, there are two roots with positive
real parts. Using a computer, we see that the roots are approximately −5.107 and 0.053 ±
1.595i. As shown by the Routh table, two of the roots have positive real parts.
7. The associated Routh table is given by

s3 1 9 0
s2 5 5 0
s 8 0
s0 5
5.9. LINEAR SYSTEMS AND FEEDBACK CONTROL 391

There are no sign changes in the first column. Therefore, there are no roots with positive
real parts. Using a computer, we see that the roots are −1 and −2 ± i. As shown by the
Routh table, none of the roots have positive real parts.
8. The associated Routh table is given by

s4 1 20 −34 0
s3 8 1 0
s2 159/8 −272/8 0
s 2335/159 0
s0 −272/8

There is one sign change in the first column. Therefore, there is one root with positive real
part. Using a computer, we see that the roots are 1.056, −2.288 and −3.384 ± 1.620i. As
shown by the Routh table, exactly one of the roots has a positive real part.
9. The associated Routh table is given by

s4 1 −5 34 0
s3 5 −35 0
s2 2 34 0
s −120 0
s0 34

There are two sign changes in the first column. Therefore, there are two roots with positive
real parts. Using a computer, we see that the roots are 1, 2 and −4 ± i. As shown by the
Routh table, exactly two of the roots have positive real parts.
10. The associated Routh table is given by

s4 1 79 221 0
s3 12 205 0
s2 743/12 221 0
s 120,491/743 0
s0 221

There are no sign changes in the first column. Therefore, there are no roots with positive
real parts. Using a computer, we see that the roots are −4.091 ± 5.083i and −1.909 ± 1.244i.
As shown by the Routh table, none of the roots have positive real parts.
11. The associated Routh table is given by

s4 1 24 100 0
s3 8 32 0
s2 20 100 0
s −8 0
s0 100
392 CHAPTER 5. THE LAPLACE TRANSFORM

There are two sign changes in the first column. Therefore, there are two roots with positive
real parts. Using a computer, we see that the roots are approximately 0.141 ± 2.141i and
−4.141 ± 2.141i. As shown by the Routh table, two of the roots have positive real parts.
12. The associated Routh table is given by
s3 1 K −6 0
s2 5 K 0
s 4K/5 − 6 0
s0 K
If we want all the poles to lie in the left half plane, we need all the entries in the first column
to have the same sign. That is, we need 4K/5 − 6 > 0 and K > 0. We see that these two
inequalities will be satisfied if and only if K > 15/2. Below we show a plot of the roots of
the polynomial s3 + 5s2 + (K − 6)s + K for K = 6, K = 15/2, and K = 9, respectively.

13. The associated Routh table is given by

s3 1 6 0
s2 4 4+K 0
s 5 − K/4 0
s0 4+K
If we want all the poles to lie in the left half plane, we need all the entries in the first column
to have the same sign. That is, we need 5 − K/4 > 0 and 4 + K > 0. We see that these two
inequalities will be satisfied if and only if −4 < K < 20. Below we show a plot of the roots
of the polynomial s3 + 4s2 + 6s + 4 + K for K = −12, K = 8, and K = 28, respectively.
5.9. LINEAR SYSTEMS AND FEEDBACK CONTROL 393

14. The associated Routh table is given by

s3 1 14 0
s2 3 8+K 0
s 34/3 − K/3 0
s0 8+K

If we want all the poles to lie in the left half plane, we need all the entries in the first column
to have the same sign. That is, we need 34 − K > 0 and 8 + K > 0. We see that these two
inequalities will be satisfied if and only if −8 < K < 34. Below we show a plot of the roots
of the polynomial s3 + 3s2 + 10s + 8 + K for K = −29, K = 13, and K = 55, respectively.

15. The associated Routh table is given by

s4 1 14 16 + K 0
s3 4 20 0
s2 9 16 + K 0
s 116/9 − 4K/9 0
s0 16 + K

If we want all the poles to lie in the left half plane, we need all the entries in the first column
to have the same sign. That is, we need 116 − 4K > 0 and 16 + K > 0. We see that these
two inequalities will be satisfied if and only if −16 < K < 29. Below we show a plot of the
roots of the polynomial s4 + 4s3 + 14s2 + 20s + 16 + K for K = −38, K = 7, and K = 51,
respectively.
394 CHAPTER 5. THE LAPLACE TRANSFORM