CHAPTER

Second Order Linear Equations

3.1
3. Let y = ert , so that y 0 = r ert and y 00 = r2 ert . Direct substitution into the
differential equation yields (6r2 − r − 1)ert = 0 . Since ert 6= 0, the characteristic
equation is 6r2 − r − 1 = 0 . The roots of the equation are r = −1/3 , 1/2 . Hence
the general solution is y = c1 e−t/3 + c2 et/2 .

5. Substitution of the assumed solution y = ert results in the characteristic equation

r2 + 5r = 0 . The roots of the equation are r = 0 , −5 . Hence the general solution
is y = c1 e0t + c2 e−5t = c1 + c2 e−5t .
√
7. The characteristic equation is r2 − 9r +√9 = 0 , with roots √
r = (9 ± 3 5)/2 .
Therefore the general solution is y = c1 e(9+3 5)t/2 + c2 e(9−3 5)t/2 .

10. Substitution of the assumed solution y = ert results in the characteristic equa-
tion r2 + 4r + 3 = 0 . The roots of the equation are r = −1 , −3 . Hence the gen-
eral solution is y = c1 e−t + c2 e−3t . Its derivative is y 0 = −c1 e−t − 3c2 e−3t . Based
on the first condition, y(0) = 2 , we require that c1 + c2 = 2 . In order to satisfy
y 0 (0) = −1 , we find that −c1 − 3c2 = −1 . Solving for the constants, c1 = 5/2
and c2 = −1/2 . Hence the specific solution is y(t) = 5e−t /2 − e−3t /2. It clearly
converges to 0 as t → ∞.
59
60 Chapter 3. Second Order Linear Equations

15. Substitution of the assumed solution y = ert results in the characteristic equa-
tion r2 + 8r − 9 = 0 . The roots of the equation are r = 1 , −9 . Hence the gen-
eral solution is y = c1 et + c2 e−9t . Its derivative is y 0 = c1 et − 9c2 e−9t . Based
on the first condition, y(1) = 1 , we require that c1 e + c2 e−9 = 1 . In order to
satisfy the condition y 0 (1) = 0 , we find that c1 e − 9c2 e−9 = 0 . Solving for the
constants, c1 = 9e−1 /10 and c2 = e9 /10. Hence the specific solution is y(t) =
9et−1 /10 + e9−9t /10 = 9e(t−1) /10 + e−9(t−1) /10. (Observe the shift on the time
axis.) It clearly increases without bound as t → ∞.

17. An algebraic equation with roots 2 and −3 is (r − 2)(r + 3) = r2 + r − 6 = 0 .

This is the characteristic equation for the differential equation y 00 + y 0 − 6 y = 0 .

19. The characteristic equation is r2 − 1 = 0, with roots r = 1, −1. Therefore the

general solution is y = c1 et + c2 e−t , with derivative y 0 = c1 et − c2 e−t . To satisfy
the initial conditions, we require that c1 + c2 = 5/4 and c1 − c2 = −3/4. Solving
for the coefficients, c1 = 1/4 and c2 = 1. This means that the specific solution is
y(t) = et /4 + e−t . From this, y 0 = et /4 − e−t = 0 when e2t = 4 or t = ln 2. The
value here is y(ln 2) = 2/4 + 1/2 = 1. Since y 00 = y is positive at t = ln 2, this is a
minimum.
3.1 61

21. The general solution is y = c1 e−t + c2 e2t . Using the initial conditions we obtain
c1 + c2 = α and −c1 + 2c2 = 2, so adding the two equations we find 3c2 = α + 2.
If y is to approach 0 as t → ∞, c2 must be zero. Thus α = −2.

24. The characteristic equation is r2 + (3 − α)r − 2(α − 1) = 0. Solving this equa-

tion, we see that the roots are r = α − 1, −2. Therefore, the general solution
is y(t) = c1 e(α−1)t + c2 e−2t . In order for all solutions to tend to zero, we need
α − 1 < 0. Therefore, the solutions will all tend to zero as long as α < 1. Due to
the term c2 e−2t , we can never guarantee that all solutions will become unbounded
as t → ∞.

25.(a) The characteristic equation is 2r2 + 3r − 2 = 0, with roots r = 1/2 and r =

−2. The initial conditions give y(t) = (2β + 1)e−2t /5 + (4 − 2β)et/2 /5.

(b) y(t) = 2 et/2 /5 + 3 e−2t /5 .

We obtain that y 0 = (−6e−2t + et/2 )/5. Setting

p this equal to zero and solving for t
yields t0 = (2 ln 6)/5. At this point, y0 = 5 3/16 ≈ 0.715485.

(c) From part (a), if β = 2 then y = e−2t and the solution simply decays to zero. For
β > 2, the solution becomes unbounded negatively, and again there is no minimum
point. For 0 < β < 2 there is always a minimum point, as found in part (b).
√
28. (a) The roots of the characteristic equation are r = (−b ± b2 − 4ac)/2a. For
the roots to be real and different we must have b2 − 4ac > 0. If they are to be
62 Chapter 3. Second Order Linear Equations

negative, then we must have b > 0 (since we are given√ that a > 0) and c > 0. This
latter condition comes from the fact that if c ≤ 0 then b2 − 4ac ≥ b and hence the
numerator of r would give both positive and negative values, or a zero if c = 0.

(b) From part (a), this will happen when b2 − 4ac > 0 and c < 0.

(c) Similarly to part (a), this happens when b2 − 4ac > 0 and b < 0 and c > 0.

3.2
2.  
 cos t sin t 
W (cos t, sin t) = 
  = cos2 t + sin2 t = 1.
− sin t cos t

4.  
x xex 
W (x, xex ) =  = xex + x2 ex − xex = x2 ex .
1 ex + xex 

8. Write the IVP as

3t 0 4 sin t
y 00 − y + y= .
t−1 t−1 t−1
Since the coefficient functions are continuous for all t < 1 and t0 = −2 < 1, the
IVP is guaranteed to have a unique solution for all t < 1. The longest interval of
existence is (−∞, 1).

12. Write the IVP as

1
y 00 + y 0 + (tan x)y = 0.
x−2
Since the coefficient functions are continuous for all x such that x 6= 2, nπ + π/2
and x0 = 3, the IVP is guaranteed to have a unique solution for all x such that
2 < x < 3π/2.

14. For y = 1, y 0 = 0 and y 00 = 0, so yy 00 + (y 0 )2 = 0. For y = t1/2 , y 0 = t−1/2 /2

and y 00 = −t−3/2 /4, thus yy 00 + (y 0 )2 = −t−1 /4 + t−1 /4 = 0. If y = c1 · 1 + c2 t1/2
is substituted into the differential equation, we get (c1 + c2 t1/2 )(−c2 t−3/2 /4) +
(c2 t−1/2 /2)2 = −c1 c2 t−3/2 /4, which is zero only if c1 = 0 or c2 = 0. Thus the linear
combination of two solutions is not, in general, a solution. Theorem 3.2.2 is not
contradicted however, since the differential equation is not linear.

15. y = φ(t) is a solution of the differential equation, so L[φ](t) = g(t). Since L

is a linear operator, L[cφ](t) = cL[φ](t) = cg(t). But, since g(t) 6= 0, cg(t) = g(t)
if and only if c = 1. This is not a contradiction of Theorem 3.2.2 since the linear
differential equation is not homogeneous.
3.2 63

18. W (t, g(t)) = tg 0 (t) − g(t) = t2 et . Dividing both sides of the equation by t, we
have g 0 − g/t = tet . This a linear equation for g with an integrating factor 1/t.
Therefore, g(t) = tet + ct.

22. The general solution is y = c1 et + c2 e−2t . W (et , e−2t ) = −3e−t , and hence
the exponentials form a fundamental set of solutions. On the other hand, the
fundamental solutions must also satisfy the conditions y1 (0) = 1, y10 (0) = 0; y2 (0) =
0, y20 (1) = 0. For y1 , the initial conditions require c1 + c2 = 1, c1 − 2c2 = 0. The
coefficients are c1 = 2/3, c2 = 1/3. For the solution y2 , the initial conditions require
c1 + c2 = 0, c1 − 2c2 = 1. The coefficients are c1 = 1/3, c2 = −1/3. Hence the
fundamental solutions are
2 1 1 1
y1 = − et + e−2t and y2 = et − e−2t .
3 3 3 3

26. For y1 = x, y10 = 1 and y100 = 0. Therefore, x2 y100 − x(x + 2)y10 + (x + 2)y1 =
−x(x + 2) + (x + 2)x = 0. For y2 = xex , y20 = (1 + x)ex and y200 = (2 + x)ex . Hence
x2 y200 − x(x + 2)y20 + (x + 2)y2 = x2 (2 + x)ex − x(x + 2)(1 + x)ex + (x + 2)xex = 0.
Further, W (x, xex ) = x2 ex 6= 0 for x > 0. Therefore, the solutions form a funda-
mental set.

28.(a) For y1 = e−t , y10 = −e−t and y100 = e−t . Therefore, y100 − y10 − 2y1 = e−t +
e−t − 2e−t = 0. For y2 = e2t , y20 = 2e2t and y200 = 4e2t . Therefore, y200 − y20 − 2y2 =
4e2t − 2e2t − 2e2t = 0. Further, W (e−t , e2t ) = 3et 6= 0 . Therefore, the functions
form a fundamental set of solutions.

(b) Since the equation is linear and y3 , y4 , y5 are all linear combinations of solutions,
they are also solutions of the differential equation by Theorem 3.2.2.

(c) W (y1 , y3 ) = −6et , W (y2 , y3 ) = 0, W (y1 , y4 ) = 6et and W (y4 , y5 ) = 0. There-

fore, {y1 , y3 } and {y1 , y4 } form fundamental sets of solutions, but {y2 , y3 } and
{y4 , y5 } do not.

29. Writing the differential Requation in the form of Eq.(22), p(t) = −(t + 2)/t. Thus
Eq.(23) yields W (t) = ce− −(t+2)/t dt = ct2 et .
R
34. From Eq.(23) we have W (y1 , y2 ) = ce− p(t) dt , where p(t) = 2/t from the differ-
ential equation. Thus W (y1 , y2 ) = c/t2 . We identify c = 2 from W (y1 , y2 )(1) = 2
and then W (y1 , y2 )(5) = 2/25.

38. Let c be the point in I at which both y1 and y2 vanish. Then W (y1 , y2 )(c) =
y1 (c)y20 (c) − y10 (c)y2 (c) = 0. Since the Wronskian is zero the functions y1 and y2
cannot form a fundamental set.

40. Suppose that y1 and y2 have a point of inflection at t0 and either p(t0 ) 6=
0 or q(t0 ) 6= 0. Since y 00 (t0 ) = 0 it follows from the differential equation that
p(t0 )y10 (t0 ) + q(t0 )y1 (t0 ) = 0 and p(t0 )y20 (t0 ) + q(t0 )y2 (t0 ) = 0. Now if p(t0 ) = 0
and q(t0 ) 6= 0, then y1 (t0 ) = y2 (t0 ) = 0, and W (y1 , y2 )(t0 ) = 0, so the solutions
cannot form a fundamental set. If p(t0 ) 6= 0 and q(t0 ) = 0, then y10 (t0 ) = y20 (t0 ) =
64 Chapter 3. Second Order Linear Equations

0, and W (y1 , y2 )(t0 ) = 0, so again the solutions cannot form a fundamental set.
If p(t0 ) 6= 0 and q(t0 ) 6= 0, then we obtain that y10 (t0 ) = −q(t0 )y1 (t0 )/p(t0 ) and
y20 (t0 ) = −q(t0 )y2 (t0 )/p(t0 ) and thus
W (y1 , y2 )(t0 ) = y1 (t0 )y20 (t0 ) − y10 (t0 )y2 (t0 ) =
= y1 (t0 )(−q(t0 )y2 (t0 )/p(t0 )) − y2 (t0 )(−q(t0 )y1 (t0 )/p(t0 )) = 0.

41. Suppose that P (x)y 00 + Q(x)y 0 + R(x)y = [P (x)y 0 ]0 + [f (x)y]0 . On expanding

the right side and equating coefficients, we find f 0 (x) = R(x) and P 0 (x) + f (x) =
Q(x). These two conditions on f can be satisfied if R(x) = Q0 (x) − P 00 (x) which
gives the necessary condition P 00 (x) − Q0 (x) + R(x) = 0.

44. P = x, Q = − cos x, R = sin x. We have P 00 − Q 0 + R = 0 . The equation is

exact. Note that [xy 0 ]0 − [(1 + cos x)y]0 = 0 . Hence xy 0
R − (1 + cos x)y = c1 . This
− (1+cos x)/x dx
equation is linear, with integrating factor µ(x) = e . Therefore the
general solution is
Z x
y(x) = [µ(x)]−1 (c1 t−1 µ(t) dt + c2 ).
x0

46. We want to choose µ(x) and f (x) so that

µ(x)P (x)y 00 + µ(x)Q(x)y 0 + µ(x)R(x)y = [µ(x)P (x)y 0 ]0 + [f (x)y]0 .
Expanding the right side and equating the coefficients of y, y 0 and y 00 gives µ0 (x)P (x) +
µ(x)P 0 (x) + f (x) = µ(x)Q(x) and f 0 (x) = µ(x)R(x). Differentiate the first equa-
tion and then eliminate f 0 (x) to obtain the adjoint equation P µ00 + (2P 0 − Q)µ0 +
(P 00 − Q0 + R)µ = 0.

48. P = 1 − x2 , Q = −2x, R = α(α + 1). Hence the coefficients are 2P 0 − Q =

−4x + 2x = −2x and P 00 − Q 0 + R = −2 + 2 + α(α + 1) = α(α + 1). The adjoint
of the original differential equation is given by (1 − x2 )µ00 − 2xµ0 + α(α + 1)µ = 0.

50. Write the adjoint as P̃ µ00 + Q̃µ0 + R̃µ = 0 where P̃ = P , Q̃ = 2P 0 − Q and

R̃ = P 00 − Q0 + R. The adjoint of this equation, namely, the adjoint of the adjoint
is P̃ y 00 + (2P̃ 0 − Q̃)y 0 + (P̃ 00 − Q̃0 + R̃)y = 0. After substitution for P̃ , Q̃ and R̃ and
simplification we obtain P y 00 + Qy 0 + Ry = 0. This is just the original equation.

51. The adjoint of P y 00 + Qy 0 + Ry = 0 is P µ00 + (2P 0 − Q)µ0 + (P 00 − Q0 + R)µ =

0. The two equations are the same if 2P 0 − Q = Q and P 00 − Q0 + R = R. This
will be true if P 0 = Q. For Problem 47, P 0 = 2x 6= x = Q, so the Bessel equation
of order ν is not self-adjoint. In a similar manner we find that the equations in
Problem 48 and 49 are self-adjoint.

3.3
1. exp (1 + 2i) = e1+2i = e1 e2i = e(cos 2 + i sin 2).
3.3 65

1−i
5. 21−i = eln 2 = e(1−i) ln 2 = eln 2 e−i ln 2 = 2(cos ln 2 − i sin ln 2).

7. The characteristic equation is r2 − 2r + 2 = 0. Therefore, the roots are r = 1 ± i.

Using Eq.(24), we arrive at the general solution y(t) = c1 et cos t + c2 et sin t.

11. The characteristic equation is r2 + 6r + 13 = 0. Therefore, the roots are r =

−3 ± 2i. Thus we arrive at the general solution y(t) = c1 e−3t cos 2t + c2 e−3t sin 2t.

14. The characteristic equation is given by 9r2 + 9r − 4 = 0. Therefore, the roots

are r = −4/3, 1/3. Therefore, the general solution is y(t) = c1 e−4t/3 + c2 et/3 .

18. The characteristic equation is r2 + 4r + 5 = 0, which has roots r = −2 ± i.

Therefore, the general solution is y(t) = c1 e−2t cos t + c2 e−2t sin t. The derivative
of y is y 0 (t) = c1 e−2t (−2 cos t − sin t) + c2 e−2t (−2 sin t + cos t). Using the initial
conditions, we have c1 = 1 and −2c1 + c2 = 0. Therefore, c1 = 1 and c2 = 2, and we
conclude that the solution is y(t) = e−2t cos t + 2e−2t sin t. The solution oscillates
as it decays to zero as t → ∞. The oscillation is hard to see on this graph, but y(t)
does cross the t axis at t = arctan(−1/2) ≈ 2.68 and periodically after that.

22. The characteristic equation is r2 + 2r + 2 = 0, which has roots r = −1 ± i.

Therefore, the general solution is y(t) = c1 e−t cos t + c2 e−t sin t. The derivative of
y is y 0 (t) = c1 e−t (− cos t − sin t) + c2 e−t (− sin t + cos t). Using the initial condi-
tions, we have

√ √
2 −π/4 2 −π/4 √
c1 e + c2 e =2 and − 2c1 e−π/4 = −2.
2 2

√ √
Therefore, c1 = 2eπ/4 and c2 = 2eπ/4 , and we conclude that the solution is

√ √
y(t) = 2e−(t−π/4) cos t + 2e−(t−π/4) sin t.

The solution oscillates as it decays to zero as t → ∞.

66 Chapter 3. Second Order Linear Equations

23.(a)
√ The characteristic equation is 3r2 − r + 2 = 0, √ which has roots r = √ 1/6 ±
i 23/6. Thus the general solution is u(t) = √ c1 et/6 cos 23t/6 + c2 et/6 sin 23t/6.
We obtain u(0) = c1 = 2 and u0 (0) = c1 /6 + 23c2 /6 = 0. Solving for c2 we find
that √ √
t/6 23 2 23
u(t) = e (2 cos t − √ sin t).
6 23 6

(b) To estimate the first time that |u(t)| = 10 plot the graph of u(t) as found in
part (a). Use this estimate in an appropriate computer software program to find
t = 10.7598.
√
25.(a) The√characteristic√ equation is r2 + 2r + 6 = 0, so r = −1 ± √ 5 i and y(t) =
e−t (c1 cos 5 t + c2 sin √ 5 t). Thus y(0)√= c1 =
√ 2 and y 0
(0) = −c 1 + 5c2 = α and
hence y(t) = e−t (2 cos 5 t + [(α + 2)/ 5] sin 5 t).

−1
√ √ √
(b) We can√see that
√ y(1) = e (2 cos 5 + [(α + 2)/ 5] sin 5) = 0, which gives
α = −2 − 2 5 cot 5 ≈ 1.508.
√ √ √ √
For y(t) = 0 we must have 2 cos √5 t + [(α + 2)/ 5]√
(c) √ sin 5 t = 0 or tan 5 t =
−2 5/(α + 2). For α ≥ 0 this yields 5t = π − arctan(2 5/(α + 2)) since arctan x
is an odd function.
√ √
(d) From part (c), arctan(2 5/(α + 2)) → 0 as α → ∞, so t → π/ 5.

31. Let r = λ + iµ, then

dert d[eλt (cos µt + i sin µt)]
= = λeλt (cos µt + i sin µt)+
dt dt
+eλt (−µ sin µt + iµ cos µt) = λeλt (cos µt + i sin µt) + iµeλt (i sin µt + cos µt) =
= eλt (λ + iµ)(cos µt + i sin µt) = rert .

33. Suppose that t = a and t = b (b > a) are consecutive zeros of y1 . We must

show that y2 vanishes once and only once in the interval a < t < b. Assume that
it does not vanish. Then we can form the quotient y1 /y2 on the interval a ≤ t ≤ b.
Note that y2 (a) 6= 0 and y2 (b) 6= 0, otherwise y1 and y2 would not be a fundamental
3.3 67

set of solutions. Next, y1 /y2 vanishes at t = a and t = b and has a derivative in

a < t < b. By Rolle’s theorem, the derivative must vanish at an interior point. But
(y1 /y2 )0 = (y10 y2 − y2 y10 )/y22 = −W (y1 , y2 )/y22 , which cannot be zero since y1 and
y2 are fundamental solutions. Hence we have a contradiction and conclude that y2
must vanish at a point between a and b. Finally, we show that it can vanish at only
one point between a and b. Suppose that it vanishes at two points c and d between
a and b. By the argument we have just given we can show that y1 must vanish
between c and d. But this contradicts the assumption that a and b are consecutive
zeros of y1 .

34.(a) Let x = ln t. We differentiate, using the Chain Rule:

dy dy dx dy 1
= =
dt dx dt dx t
and
d2 y d2 y 1
     
d dy 1 dy 1 dy 1
= + − = + − 2 .
dt2 dt dx t dx t2 dx2 t2 dx t

(b) Using part (a), we can see that

d2 y dy
t2 + αt + βy = 0
dt2 dt
transforms into
d2 y dy dy d2 y dy
2
− +α + βy = 2 + (α − 1) + βy = 0.
dx dx dx dx dx

36. The equation transforms into y 00 + 3y 0 + 2y = 0. The characteristic roots are

r = −1, −2. The solution is
c1 c2
y = c1 e−x + c2 e−2x = c1 e− ln t + c2 e−2 ln t = + 2.
t t

40. The equation transforms into y 00 − 2y 0 + 5y = 0. The characteristic roots are

r = 1 ± 2i. The solution is
y = c1 ex cos(2x) + c2 ex sin(2x) = c1 t cos(2 ln t) + c2 t sin(2 ln t).

2
44. We use the result of Problem 43. Note that p(t) = t and q(t) = e−t > 0 for
−∞ < t < ∞. Thus (q 0 + 2pq)/q 3/2 = 0 and the differential equation can be trans-
2
formed into an equation with constant coefficients by letting x = u(t) = e−t /2 dt.
R

Substituting x = u(t) in the differential equation found in part (b) of Problem 43

we obtain, after dividing by the coefficient of d2 y/dx2 , the differential equation
d2 y/dx2 + y = 0. Hence the general solution of the original differential equation is
2
y(t) = c1 cos x(t) + c2 sin x(t), where x(t) = e−t /2 dt.
R
68 Chapter 3. Second Order Linear Equations

3.4
1. The characteristic equation is given by r2 − 2r + 1 = 0. Therefore, we have one
repeated root r = 1, and the general solution is given by y(t) = c1 et + c2 tet .

5. The characteristic equation is r2 − 2r + 10 = 0 , with complex roots r = 1 ± 3i.

The general solution is y(t) = c1 et cos 3t + c2 et sin 3t .

9. The characteristic equation is given by 25r2 − 20r + 4 = 0. Therefore, we have

one repeated root r = 2/5, and the general solution is given by y(t) = c1 e2t/5 +
c2 te2t/5 .

12. The characteristic equation is given by r2 − 6r + 9 = 0. Therefore, there is one

repeated root, r = 3, and the general solution is given by y(t) = c1 e3t + c2 te3t . Af-
ter differentiation, y 0 (t) = 3c1 e3t + c2 (1 + 3t)e3t . Now using the initial conditions,
we need c1 = 0 and 3c1 + c2 = 2. The solution of this system of equations is c1 = 0
and c2 = 2, and the specific solution is y(t) = 2te3t . The solution y → ∞ as t → ∞.

14. The characteristic equation is given by r2 + 4r + 4 = 0. Therefore, there is one

repeated root, r = −2, and the general solution is given by y(t) = c1 e−2t + c2 te−2t .
We can see that y 0 (t) = −2c1 e−2t + c2 (1 − 2t)e−2t . Now using the initial conditions,
we need c1 e2 − c2 e2 = 2 and −2c1 e2 + 3c2 e2 = 1. The solution of this system of
equations is c1 = 7e−2 and c2 = 5e−2 , and the specific solution is y(t) = 7e−2(t+1) +
5te−2(t+1) . The solution y → 0 as t → ∞.
3.4 69

17.(a) The characteristic equation is 4r2 + 4r + 1 = (2r + 1)2 = 0, so we have y(t) =

(c1 + c2 t)e−t/2 . Thus y(0) = c1 = 1 and y 0 (0) = −c1 /2 + c2 = 0 and hence c2 = 5/2
and y(t) = (1 + 5t/2)e−t/2 .

(b) From part (a), y 0 = −(1/2)(1 + 5t/2)e−t/2 + (5/2)e−t/2 = 0, when −1/2 − 5t/4 +
5/2 = 0, so tM = 8/5 and yM = 5e−4/5 .

(c) From part (a), c1 is the same and y 0 (0) = −1/2 + c2 = b or c2 = b + 1/2 and
y(t) = [1 + (b + 1/2)t]e−t/2 .

(d) From part (c), y 0 = −(1/2)[1 + (b + 1/2)t]e−t/2 + (b + 1/2)e−t/2 = 0 which yields

tM = 4b/(2b + 1) → 2 as b → ∞ and yM = (1 + (2b + 1)/2 · 4b/(2b + 1))e−2b/(2b+1) =
(1 + 2b)e−2b/(2b+1) . Since e−2b/(2b+1) → e−1 as b → ∞, yM → ∞ as b → ∞.

19. Suppose the roots are distinct, r1 < r2 . Then the solution is y(t) = c1 er1 t +
c2 er2 t . Solving the equation y(t) = 0, we see that we must have c1 er1 t = −c2 er2 t
which implies e(r1 −r2 )t = −c2 /c1 . First, in order to guarantee any solution of this
equation, we would need c2 /c1 < 0. Then, applying the natural logarithm function
to the equation, we see that t = ln(−c2 /c1 )/(r1 − r2 ).
If the roots are not distinct, then the solution is given by y(t) = c1 ert + c2 tert .
Therefore, y(t) = 0 implies (c1 + c2 t)ert = 0. Since ert 6= 0, we must have c1 + c2 t =
0. Therefore, the solution will be zero only when t = −c1 /c2 .

21. If r1 6= r2 , then φ(t; r1 , r2 ) = (er2 t − er1 t )/(r2 − r1 ) is defined for all t. Note
that φ is a linear combination of the fundamental solutions, er1 t and er2 t , so φ
is a solution of the differential equation. The limit of φ as r2 → r1 is (by its
definition) the derivative of the function ert with respect to r at the point r1 , hence
φ(t; r1 , r2 ) → ter1 t as r2 → r1 .

25. Following the reduction of order technique given, y1 = 1/t, p(t) = 3/t, so the
equation for v is v 00 /t + v 0 /t2 = 0. After separating the variables the equation
becomes v 00 /v 0 = −1/t, so ln v 0 = − ln t + c. We obtain that v 0 = c/t and then v =
c ln t. Thus the second solution is y2 = ln t/t.

27. Following the reduction of order technique given, y1 = sin(x2 ), p(x) = −1/x,
so the equation for v is sin(x2 )v 00 + (4x cos(x2 ) − sin(x2 )/x)v 0 = 0. After separat-
ing the variables the equation becomes v 00 /v 0 = 1/x − 4x cot(x2 ), so ln v 0 = ln x −
2 ln(sin(x2 )). We obtain that v 0 = x/ sin2 (x2 ) and then v = − cot(x2 )/2. Thus the
second solution is y2 = cos(x2 ).

30. Let y2 (x) = [x−1/2 sin x]v(x). Substituting y2 into the differential equation, we
conclude that sin x v 00 + 2 cos x v 0 = 0. This equation is linear in v 0 . Its solution is
given by v 0 (x) = c/ sin2 x. Integrating with respect to x, we have v(x) = c1 cot x +
c2 . Therefore, y2 (x) = c1 x−1/2 cos x + c2 x−1/2 sin x. Since we already have the
solution y1 (x) = x−1/2 sin x, we take the solution y2 (x) = x−1/2 cos x.

32. R (y2 /y1 )0 = (y20 y1 − y10 y2 )/y12 = W (y1R, y2 )/y12 . Abel’s formula is W (y1 , y2 ) =
− t p(r) dr − t p(r) dr
ce t0 . Hence (y2 /y1 )0 = cy1−2 e t0 . Integrating and setting c = 1
70 Chapter 3. Second Order Linear Equations

(since a solution y2 can be multiplied by any constant) and taking the constant
R t − R s p(r) dr 2
of integration to be zero we obtain y2 = y1 t0 [e s0
/y1 (s)] ds.
R
34. From Problem 32 and Abel’s formula we have (y2 /y1 )0 = e (1/t) dt / sin2 t2 =
eln t / sin2 t2 = t csc2 t2 . Thus y2 /y1 = −(1/2) cot t2 and hence we can choose y2 =
cos t2 since y1 = sin t2 .

37. The general √ solution of the differential equation is y = c1 er1 t + c2 er2 t where
2 2
r1 , r2 = (−b ± b − 4ac)/2a, provided√ b − 4ac 6= 0. In this case there are two
possibilities. If b2 − 4ac > 0 then b2 − 4ac < b and r1 and r2 are real and nega-
tive. Consequently, y → 0 as t → ∞. If b2 − 4ac < 0, then r1 and r2 are complex
conjugates with negative real part. Again, y → 0 as t → ∞. Finally, if b2 − 4ac = 0,
then y = c1 ert + c2 tert where r = −b/2a < 0. Hence, again y → 0 as t → ∞. This
conclusion does not hold if either b = 0 (since in this case y(t) = c1 cos ωt + c2 sin ωt,
where ω 2 = c/a) or c = 0 (since one of the solutions would be y1 (t) = c1 ).

41. Letting x = ln t, the equation transforms into y 00 + y 0 + (1/4)y = 0. We get a

repeated root r = −1/2, so the solution is y = c1 e−x/2 + c2 xe−x/2 = c1 e− ln t/2 +
c2 ln te− ln t/2 = c1 t−1/2 + c2 t−1/2 ln t.

45. Letting x = ln t, the equation transforms into y 00 + 4y 0 + 13y = 0. The charac-

teristic roots are r = −2 ± 3i. The solution is y = c1 e−2x cos(3x) + c2 e−2x sin(3x) =
c1 t−2 cos(3 ln t) + c2 t−2 sin(3 ln t).

3.5
1. The characteristic equation for the homogeneous problem is r2 − 2r − 3 = 0,
which has roots r = 3, −1. Therefore, the solution of the homogeneous problem is
yh (t) = c1 e3t + c2 e−t . To find a solution of the nonhomogeneous problem, we look
for a solution of the form yp (t) = Ae2t . Substituting a function of this form into
the differential equation, we have 4Ae2t − 4Ae2t − 3Ae2t = 3e2t . This means that
we need −3A = 3, or A = −1. Hence the general solution of the nonhomogeneous
problem is y(t) = c1 e3t + c2 e−t − e2t .

3. The characteristic equation for the homogeneous problem is r2 − r − 2 = 0 ,

with roots r = −1, 2. Hence yh (t) = c1 e−t + c2 e2t . Set now yp = At2 + Bt + C.
Substitution into the given differential equation, and comparing the coefficients,
results in the system of equations −2A = 4, −2A − 2B = −2 and 2A − B − 2C = 0.
Hence yp = −2t2 + 3t − 7/2. The general solution is y(t) = yh (t) + yp (t).

6. The characteristic equation for the homogeneous problem is r2 + 2r = 0, which

has roots r = 0, −2. Therefore, the solution of the homogeneous problem is yh (t) =
c1 + c2 e−2t . To find a solution of the nonhomogeneous problem, we look for a
solution of the form yp (t) = At + B cos 2t + C sin 2t. Substituting a function of
this form into the differential equation, and equating like terms, we have 2A = 3,
−4B + 4C = 0 and −4B − 4C = 4. The solution of these equations is A = 3/2,
3.5 71

B = −1/2 and C = −1/2. Therefore, the general solution of the nonhomogeneous

problem is y(t) = c1 + c2 e−2t + 3t/2 − cos 2t/2 − sin 2t/2.

8. The characteristic equation for the homogeneous problem is r2 + 2r + 1 = 0,

which has the repeated root r = −1. Therefore, the solution of the homogeneous
problem is yh (t) = c1 e−t + c2 te−t . To find a solution of the nonhomogeneous prob-
lem, we look for a solution of the form yp (t) = At2 e−t . Substituting a func-
tion of this form into the differential equation, and equating like terms, we have
2A = 2. Therefore, A = 1 and the general solution of the nonhomogeneous problem
is y(t) = c1 e−t + c2 te−t + t2 e−t .

10. The characteristic equation for the homogeneous problem is r2 + 1 = 0, which

has roots r = ±i. Therefore, the solution of the homogeneous problem is yh (t) =
c1 cos t + c2 sin t. To find a solution of the nonhomogeneous problem, we look for
a solution of the form yp (t) = A cos 2t + B sin 2t + Ct cos 2t + Dt sin 2t. Substitut-
ing a function of this form into the differential equation, and equating like terms,
we have 4D − 3A = 0, −3B − 4C = 3, −3C = 1, and −3D = 0. The solution of
these equations is A = 0, B = −5/9, C = −1/3 and D = 0. Therefore, the genaral
solution of the nonhomogeneous problem is y(t) = c1 cos t + c2 sin t − 5 sin 2t/9 −
t cos 2t/3.

2
13. The characteristic equation√ for the homogeneous problem is r + r + 4 = 0,
which has roots r = (−1 ± i 15)/2.√ Therefore,√the solution of the homogeneous
problem is yh (t) = e−t/2 (c1 cos( 15t/2) + c2 sin( 15t/2)). To find a solution of the
nonhomogeneous problem, we look for a solution of the form yp (t) = Aet + Be−t .
Substituting a function of this form into the differential equation, and equating like
terms, we have 6A = 1 and 4B = −1. The solution of these equations is A = 1/6
and B = −1/4. Therefore,√ the general√solution of the nonhomogeneous problem
is y(t) = e−t/2 (c1 cos( 15t/2) + c2 sin( 15t/2)) + et /6 − e−t /4. In this case, we
could also assume that yp is a linear combination of sinh t and cosh t.

15. The characteristic equation for the homogeneous problem is r2 + r − 2 = 0,

which has roots r = 1, −2. Therefore, the solution of the homogeneous problem is
yh (t) = c1 et + c2 e−2t . To find a solution of the nonhomogeneous problem, we look
for a solution of the form yp (t) = At + B. Substituting a function of this form into
the differential equation, and equating like terms, we have −2A = 2 and A − 2B =
0. The solution of these equations is A = −1 and B = −1/2. Therefore, the general
solution of the nonhomogeneous problem is y(t) = c1 et + c2 e−2t − t − 1/2. The
initial conditions imply c1 + c2 − 1/2 = 0 and c1 − 2c2 − 1 = 1. Therefore, c1 = 1
and c2 = −1/2 which implies that the solution of the IVP is y(t) = et − e−2t /2 −
t − 1/2.

18. The characteristic equation for the homogeneous problem is r2 − 2r − 3 = 0,

which has roots r = 3, −1. Therefore, the solution of the homogeneous problem
is yh (t) = c1 e3t + c2 e−t . To find a solution of the nonhomogeneous problem, we
look for a solution of the form yp (t) = Ae2t + Bte2t . Substituting a function of
this form into the differential equation, and equating like terms, we have yp (t) =
−2e2t /3 − te2t . Therefore, the general solution of the nonhomogeneous problem is
72 Chapter 3. Second Order Linear Equations

y(t) = c1 e3t + c2 e−t − 2e2t /3 − te2t . The initial conditions imply c1 + c2 − 2/3 = 1
and 3c1 − c2 − 4/3 − 1 = 0. Therefore, c1 = 1 and c2 = 2/3 which implies that the
solution of the IVP is y(t) = e3t + 2e−t /3 − 2e2t /3 − te2t .

21.(a) The characteristic equation for the homogeneous problem is r2 + 3r = 0,

which has roots r = 0, −3. Therefore, the solution of the homogeneous problem is
yh (t) = c1 + c2 e−3t . After inspection of the nonhomogeneous term, for 2t4 we must
assume a fourth order polynomial, for t2 e−3t we must assume a quadratic polyno-
mial times the exponential, and for sin 3t we must assume C sin 3t + D cos 3t. How-
ever, since e−3t and a constant are solutions of the homogeneous differential equa-
tion, we must multiply the coefficient of e−3t and the polynomial by t. The correct
form then is Y (t) = t(A0 t4 + A1 t3 + A2 t2 + A3 t + A4 ) + t(B0 t2 + B1 t + B2 )e−3t +
C sin 3t + D cos 3t.

(b) Substituting a function of this form into the differential equation, and equating
like terms, we have A0 = 2/15, A1 = −2/9, A2 = 8/27, A3 = −8/27, A4 = 16/81,
B0 = −1/9, B1 = −1/9, B2 = −2/27, C = −1/18, D = −1/18 Therefore, the gen-
eral solution of the nonhomogeneous problem is
2 4 2 3 8 8 16
y(t) = c1 + c2 e−3t + t( t − t + t2 − t + )+
15 9 27 27 81
1 1 2 sin 3t cos 3t
+t(− t2 − t − )e−3t − − .
9 9 27 18 18

24.(a) The characteristic equation for the homogeneous problem is r2 + 2r + 2 = 0,

which has roots r = −1 ± i. Therefore, the solution of the homogeneous problem is
yh (t) = e−t (c1 cos t + c2 sin t). After inspection of the nonhomogeneous term, since
e−t cos t and e−t sin t are solutions of the homogeneous differential equation, it is
necessary to multiply again by t in the particular solution, so the desired form is
Y (t) = Ae−t + t(B0 t2 + B1 t + B2 )e−t cos t + t(C0 t2 + C1 t + C2 )e−t sin t.

(b) Substituting a function of this form into the differential equation, and equating
like terms, we have A = 3, B0 = −2/3, B1 = 0, B2 = 1, C0 = 0, C1 = 1, C2 = 1.
Therefore, the general solution of the nonhomogeneous problem becomes y(t) =
e−t (c1 cos t + c2 sin t) + 3e−t + t((−2/3)t2 + 1)e−t cos t + t(t + 1)e−t sin t.

29.(a) For Y = ve−t , we have Y 0 = v 0 e−t − ve−t and Y 00 = v 00 e−t − 2v 0 e−t + ve−t .
Then Y 00 − 3Y 0 − 4Y = 2e−t implies that v 00 e−t − 2v 0 e−t + ve−t − 3v 0 e−t + 3ve−t −
4ve−t = 2e−t . Simplifying this equation, we have v 00 − 5v 0 = 2.

(b) We see that (v 0 )0 − 5(v 0 ) = 2. Therefore, letting w = v 0 , we see that w must

satisfy w0 − 5w = 2. This equation is linear with integrating factor µ(t) = e−5t .
Therefore, we have [e−5t w]0 = 2e−5t which implies that w = −2/5 + ce5t .

(c) Integrating w, we see that v = (−2/5)t + c1 e5t /5 + c2 . Then using the fact that
Y = ve−t , we conclude that Y (t) = −2te−t /5 + c1 e4t /5 + c2 e−t . Here the first term
is yp and the last two terms comprise yh .
3.5 73

31. The solution of the homogeneous problem is yh (t) = c1 cos t + c2 sin t. To solve
the nonhomogeneous IVP (starting at t = 0), we begin by looking for a solution of
the nonhomogeneous equation of the form Y = A + Bt. Substituting a function of
this form into the ODE leads to the equation A + Bt = t. Therefore, A = 0 and B =
1, and the solution of the nonhomogeneous problem is y(t) = c1 cos t + c2 sin t + t.
Now, we consider the initial conditions. The initial conditions imply that c1 = 0 and
c2 + 1 = 1. Therefore, the solution of the nonhomogeneous problem for 0 ≤ t ≤ π is
given by y1 (t) = t. Now we need to solve the nonhomogeneous problem starting at
t = π. We look for a solution of the nonhomogeneous problem of the form Y (t) =
Ce−t . Substituting a function of this form into the ODE leads to the equation
Ae−t + Ae−t = πeπ−t . Therefore, 2A = πeπ , or A = πeπ /2. Therefore, a solution
of the nonhomogeneous problem (starting at t = π) is given by y(t) = d1 cos t +
d2 sin t + πeπ−t /2. Using the solution of the IVP starting at t = 0, y1 (t) = t, we
see that at time t = π, y1 (π) = π and y10 (π) = 1. Using these as our new initial
conditions for t = π, we see that d1 , d2 must satisfy −d1 + π/2 = π and −d2 −
π/2 = 1. The solution of these equations is d1 = −π/2, d2 = −1 − π/2. Therefore,
we conclude that the solution of the nonhomogeneous IVP is

t 0≤t≤π
y(t) =
− π2 cos t − 1 + π2 sin t + π2 eπ−t


t > π.

(a) The nonhomogeneous term (b) The solution y(t)

33. According to Theorem 3.5.1, the difference of any two solutions of the lin-
ear second order nonhomogeneous differential equation is a solution of the cor-
responding homogeneous differential equation. Hence Y1 − Y2 is a solution of
ay 00 + by 0 + cy = 0. In Problem 37 of Section 3.4 we showed that if a > 0, b > 0
and c > 0 then every solution of this differential equation goes to zero as t → ∞.
If b = 0, then yc involves only sines and cosines, so Y1 − Y2 does not approach zero
as t → ∞.

36. We have (D2 − 3D − 4)y = (D − 4)(D + 1)y . Let u = (D + 1)y , and consider
the ODE u 0 − 4u = 3e2t . The general solution is u(t) = −3e2t /2 + c e4t . We there-
fore have the first order equation y 0 + y = −3e2t /2 + c e4t . The general solution of
the latter differential equation is y(t) = −e2t /2 + c1 e4t + c2 e−t .
74 Chapter 3. Second Order Linear Equations

3.6
2. Two linearly independent solutions of the homogeneous differential equation
are y1 (t) = e2t and y2 (t) = e−t . Assume Y = u1 (t)e2t + u2 (t)e−t , then Y 0 (t) =
[2u1 (t)e2t − u2 (t)e−t ] + [u01 (t)e2t + u02 (t)e−t ]. We set u01 (t)e2t + u02 (t)e−t = 0. Then
Y 00 (t) = 4u1 e2t + u2 e−t + 2u01 e2t − u02 e−t and substituting into the differential equa-
tion gives 2u01 (t)e2t − u02 (t)e−t = 2e−t (the terms involving u1 and u2 add to zero
since e−t and e2t are solutions of the homogeneous equation). Thus we have
two algebraic equations for u01 (t) and u02 (t) with the solution u01 (t) = 2e−3t /3 and
u02 (t) = −2/3. Hence u1 (t) = −2e−3t /9 and u2 (t) = −2t/3. Substituting in the ex-
pression for Y (t) we obtain Y (t) = −2e−t /9 − 2te−t /3. Since e−t is a solution of
the homogeneous differential equation, we can choose Y (t) = −2te−t /3.

5. Since cos t and sin t are solutions of the homogeneous differential equation,
we assume Y (t) = u1 (t) cos t + u2 (t) sin t. Thus Y 0 = −u1 (t) sin t + u2 (t) cos t, after
setting u01 (t) cos t + u02 (t) sin t = 0. Finding Y 00 and substituting into the differential
equation then yields −u01 (t) sin t + u02 (t) cos t = tan t. The two equations for u01 (t)
and u02 (t) have the solution u01 (t) = − sin2 t/ cos t = − sec t + cos t and u02 (t) = sin t.
Thus u1 (t) = sin t − ln(sec t + tan t) and u2 (t) = − cos t, which when substituted
into the assumed form for Y , simplified, and added to the homogeneous solution
yields y(t) = c1 cos t + c2 sin t − cos t ln(sec t + tan t).

11. The solution of the homogeneous equation is yh (t) = c1 e3t + c2 e2t . The func-
tions y1 (t) = e3t and y2 (t) = e2t form a fundamental set of solutions. The Wron-
skian of these functions is W (y1 , y2 ) = −e5t . Using the method of variation of
parameters, a particular solution is given by Y (t) = u1 (t)y1 (t) + u2 (t)y2 (t) where
Z 2s Z
e (g(s))
u1 (t) = − ds = e−3s g(s)ds
W (s)
Z 3s Z
e (g(s))
u2 (t) = ds = − e−2s g(s)ds.
W (s)
Therefore, a particular solution is Y (t) = e3(t−s)Rg(s)ds − e2(t−s)Rg(s)ds. There-
R R

fore, the general solution is y(t) = c1 e3t + c2 e2t + e3(t−s) g(s)ds − e2(t−s) g(s)ds.

14. By direct substitution, it can be verified that y1 (t) = t and y2 (t) = tet are solu-
tions of the homogeneous equation. The Wronskian of these functions is W (y1 , y2 ) =
t2 et . Rewriting the equation in standard form, we have
t(t + 2) 0 t + 2
y 00 − y + 2 y = 2t.
t2 t
Therefore, g(t) = 2t. Using the method of variation of parameters, a particular
solution is given by Y (t) = u1 (t)y1 (t) + u2 (t)y2 (t) where
tet (2t)
Z Z
t(2t)
u1 (t) = − dt = −2t and u2 (t) = dt = −2e−t .
W (t) W (t)
3.6 75

Therefore, a particular solution is Y (t) = −2t2 − 2t. (Since t is a solution of the

homogeneous differential equation, we can choose only the −2t2 part.)

18. By direct substitution, it can be verified that y1 (x) = x−1/2 sin x and y2 (x) =
x−1/2 cos x are solutions of the homogeneous equations. The Wronskian of these
functions is W (y1 , y2 ) = −1/x. Rewriting the equation in standard form, we have
1 0 x2 − 0.25 sin x
y 00 + y + y = 3 1/2 .
x x2 x
Therefore, g(x) = 3x−1/2 sin x. Using the method of variation of parameters, a
particular solution is given by Y (x) = u1 (x)y1 (x) + u2 (x)y2 (x) where
Z −1/2
x cos x(3x−1/2 sin x) 3
u1 (x) = − dx = − cos2 x
W (x) 2
Z −1/2
x sin x(3x−1/2 sin x) 3 3x
u2 (x) = dx = cos x sin x − .
W (x) 2 2
Therefore, a particular solution is
 
3 −1/2 3 3x 3
Y (x) = − x 2
cos x sin x + cos x sin x − x−1/2 cos x = − x1/2 cos x.
2 2 2 2

22. Equation (28) is

Z t Z t
y2 (s)g(s) y1 (s)g(s)
Y (t) = −y1 (t) ds + y2 (t) ds,
t0 W (y1 , y2 )(s) t0 W (y1 , y2 )(s)
where t0 is now considered the initial point. Bringing the terms y1 (t) and y2 (t)
inside the integrals and using the fact that W (y1 , y2 )(s) = y1 (s)y20 (s) − y10 (s)y2 (s),
the desired result holds. To show that Y (t) satisfies L[y] = g(t) we must take
Rt
the derivative using Leibniz’s rule, which says that if y(t) = t0 G(t, s) ds, then
Rt
Y 0 (t) = G(t, t) + t0 Gt (t, s) ds. Letting G(t, s) be the above integrand, we have
that G(t, t) = 0 and
∂G y1 (s)y20 (t) − y10 (t)y2 (s)
= g(s).
∂t W (y1 , y2 )(s)
Likewise,
t t
∂2G y1 (s)y200 (t) − y100 (t)y2 (s)
Z Z
00 ∂G(t, t)
Y = + (t, s) ds = g(t) + g(s) ds.
∂t t0 ∂t2 t0 W (y1 , y2 )(s)
Since y1 and y2 are solutions of L[y] = 0, we have L[Y ] = g(t) since all the terms
involving the integral will add to zero. Clearly y(t0 ) = 0 and y 0 (t0 ) = 0.

25. The given linear operator L[y] = [D2 − 2λD + (λ2 + µ2 )]y can be written as
L[y] = y 00 − 2λy 0 + (λ2 + µ2 )y. Therefore, p(t) = −2λ and q(t) = λ2 + µ2 . To solve
the given nonhomogeneous problem, we first need to solve the associated homoge-
neous problem. In particular, we need to look for the general solution of
y 00 − 2λy 0 + (λ2 + µ2 )y = 0.
76 Chapter 3. Second Order Linear Equations

The roots of the associated equation are λ ± iµ. Therefore, the general solu-
tion of the homogeneous problem is y(t) = eλt (c1 cos(µt) + c2 sin(µt)). Letting
y1 (t) = eλt cos(µt) and y2 (t) = eλt sin(µt), we now use the result in Problem 22.
In particular, a solution of the indicated problem will be given by
Z t λs
e cos(µs)eλt sin(µt) − eλt cos(µt)eλs sin(µs)
Y (t) = g(s) ds
t0 µe2λs
1 t λ(t−s)
Z
= e sin(µ(t − s))g(s)ds.
µ t0

29. First we put the differential equation in standard form by dividing by t2 ; we

obtain y 00 − 2y 0 /t + 2y/t2 = 4. Assuming that y = tv(t) and substituting Rin the dif-
ferential equation we obtain tv 00 = 4. Hence v 0 = 4 ln t + c2 and v(t) = 4 ln t dt +
c2 t + c1 = 4(t ln t − t) + c2 t + c1 , using integration by parts. Thus y = 4t2 ln t +
c3 t2 + c1 t, where c3 = c2 − 4. Since y1 = c1 t, we can take y2 = 4t2 ln t + c3 t2 , where
c3 t2 represents the second fundamental solution of the related homogeneous equa-
tion and 4t2 ln t is the particular solution.

3.7
√ √
2. R cos δ = −1 and R sin δ = 3 implies R = 4 = 2 and δ = π + arctan(−3) =
2π/3. Note that we have to add π to the inverse tangent value since δ must be a
second quadrant angle. Therefore, u = 2 cos(t − 2π/3).

6. The spring constant is k = .98/.05 = 19.6 N/m. The mass m = 0.1 kg. There-
fore, the equation of motion is 0.1u00 + 19.6u = 0, which can be simplified to u00 +
196u = 0. The initial conditions are u(0) = 0 cm, u0 (0) = 10 cm/sec. The general
solution of the differential equation is u(t) = A cos 14t + B sin 14t. The initial con-
dition implies A = 0 and B = 5/7. Therefore, the solution is u(t) = (5/7) sin 14t
cm. The period is T = 2π/14 seconds. Therefore, the mass will first return to its
equilibrium position in half that time; that is, in π/14 seconds.

8. The inductance L = 1 henry. The resistance R = 0. The capacitance C =

0.25 × 10−6 farads. Therefore, the equation for charge Q is Q00 + (4 × 106 )Q = 0.
The initial conditions are Q(0) = 10−6 coulombs, Q0 (0) = 0 coulombs/sec. The
general solution of this equation is Q(t) = A cos(2000t) + B sin(2000t). The initial
conditions imply the specific solution is Q(t) = 10−6 cos(2000t) coulombs.

9. The spring constant is k = .196/.05 = 3.92 N/m. The mass m = .02 kg. The
damping constant is γ = 400 dyne-sec/cm = .4 N-sec/m. Therefore, the equation
of motion is .02u00 + .4u0 + 3.92u = 0 or u00 + 20u0 + 196u = 0, with initial con-
0
ditions u(0) = √.02 m, u (0)√ = 0 m/sec. The solution of this equation is u(t) =
−10t
e (A cos(4 6t) + B sin(4 6t)). The initial conditions imply that
h √ √ √ i
y(t) = e−10t 2 cos(4 6t) + (5/ 6) sin(4 6t) cm
3.7 77

√ √
The quasi frequency is ν = 4 6 rad/sec. The quasi period is Td = π 6/12. For
undamped motion, the equation would be y 00 + 196y = 0, which has the general
solution y(t) = A cos(14t) + B sin(14t). Therefore, the period for the undamped
motion is T = π/7.√The ratio of the quasi period to the period of the undamped
motion is Td /T = 7 6/12 ≈ 1.4289. Using a computer software program we obtain
that the solution y will satisfy |y(τ )| < 0.05 for all τ > .4045 seconds.

12. The inductance L = 0.2 henry. The resistance R = 3 × 102 ohms. The capac-
itance C = 10−5 farads. Therefore, the equation for charge Q is 0.2Q00 + 300Q0 +
105 Q = 0, which can be rewritten as Q00 + 1500Q0 + 500, 000Q = 0. The initial
conditions are Q(0) = 10−6 coulombs, Q0 (0) = 0 couloumbs/sec. The general so-
lution of the differential equation is Q(t) = Ae−500t + Be−1000t . The initial condi-
tions imply A + B = 10−6 and −500A − 1000B = 0. Therefore, A = 2 × 10−6 and
B = −10−6 and the specific solution is Q(t) = 10−6 (2e−500t − e−1000t ).

17. The spring constant is k = 8/(1.5/12) = 64 lb/ft. The mass is m = 8/32 = 1/4
lb-s2 /ft. Therefore, the equation of motion is u00 /4 + γu0 + 64u = 0. Using the
quadratic√ formula, the motion will experience critical damping when γ 2 − 4km = 0,
i.e. γ = 2 km = 8 lb-s/ft.
√
√ If the system is critically damped or overdamped, then γ ≥ 2 km.
19. If γ =
2 km (critically damped), then the solution is given by u(t) = (A + Bt)e−γt/2m . In
this case, if u = 0, then we must have A + Bt = 0, that is, t = −A/B √ (assuming B 6=
0). If B = 0, the solution is never zero (unless A = 0). If γ > 2 km (overdamped),
then the solution is given by u(t) = Aer1 t + Ber2 t , where r1 , r2 are given by equation
(22) in the text. Assume for the moment, that A, B 6= 0. Then u = 0 implies
Aer1 t = −Ber2 t which implies e(r1 −r2 )t = −B/A. There is only one solution to this
equation. If A = 0 or B = 0, then there are no solutions to the equation u = 0
(unless they are both zero, in which case, the solution is identically zero).

20. If the system is critically damped, then the general solution is u(t) = (A +
Bt)e−γt/2m . The initial conditions imply A = u0 and B = v0 + (γu0 /2m). If
v0 = 0, then B = γu0 /2m. In this case, the specific solution is u(t) = u0 (1 +
γt/2m)e−γt/2m . As t → ∞, u → 0. In order for u = 0, we would need 1 + γt/2m =
0, but there are no positive times t satisfying this equation. If v0 6= 0, the specific
solution is u(t) = (u0 + (v0 + (γu0 /2m))t)e−γt/2m . In order to find a positive time
such that u(t) = 0, we need u0 + (v0 + (γu0 /2m))t = 0. The solution of this equa-
78 Chapter 3. Second Order Linear Equations

tion is t = −u0 /[v0 + (γu0 /2m)]. In order for there to be a positive time t satisfying
this equation, we need v0 + (γu0 /2m) < 0. That is, we need v0 < −γu0 /2m.

23. For Problem 17, the mass is m = 1/4 lb-s2 /ft. We suppose that ∆ = 3 and
Td = 0.3 seconds. From Problem 21, we know that ∆ = γTd /2m. Therefore, we
have 3 = 0.3γ/[2(1/4)] = 0.6γ. Therefore, the damping coefficient is γ = 5 lb-sec/ft.
p p
24. The general solution of this equation
p is u(t) = A cos( 2k/3 t) + B sin( 2k/3 t).
The period of this system is T = 2π/p 2k/3 sec. Therefore, if the period is T =
π seconds, then k must satisfy 2π/ 2k/3 = π. The solution of this equation
0
is
p k = 6. The initial conditions u(0) = 2 and u (0) = v imply A = 2 and B =
2k/3 = v. Since
p k = 6, we know that 2B = v. The amplitude of the system
√
is Ap 2 + B2 = 22 + (v/2)2 . Since the amplitude is assumed to be 3, we must
√
2 2
have 2 + (v/2) = 3 which implies v = ±2 5.

27. First, consider the static case (which is the equilibrium position). Let ∆l de-
note the length of the block below the surface of the water. The weight of the
block, which is a downward force, is w = ρl3 g. This is balanced by an equal and
opposite buoyancy force B, which is equal to the weight of the displaced water.
Thus B = ρ0 l2 ∆lg = ρl3 g. Now let u(t) be the displacement of the block from its
equilibrium position. We take downward as the positive direction. In a displaced
position the forces acting on the block are its weight, which acts downward and is
unchanged, and the buoyancy force which is now ρ0 l2 (∆l + u)g and acts upward.
The resultant force must be equal to the mass of the block times the accelera-
tion, namely ρl3 u00 . Hence ρl3 g − ρ0 l2 (∆l + u)g = ρl3 u00 . Hence the differential
equation for the motion of the block is ρl3 u00 + ρ0 l2 gu = 0 or u00 + ρ0 g/(ρl)u = 0.
This gives a simple harmonic motion with frequency (ρ0 g/ρl)1/2 and natural period
T = 2π(ρl/ρ0 g)1/2 .
√
29.(a) The characteristic√equation is 4r2 √+ r + 8 = 0, so r = (−1 ± 127)/8 and
hence u(t) = e−t/8 (c1 cos 127t/8 + c2 sin 127t/8). The initial condition u(0) √=0
0
implies that c1 =√0 and the initial
√ condition u (0) = 2 implies that c2 = 16/ 127.
Thus u(t) = (16/ 127)e−t/8 sin 127t/8.

(b)
3.7 79

(c)

The direction of motion is clockwise since the graph starts at (0, 2) and u increases
initially.

30.(a) The kinetic energy is given by mv 2 /2, where v is velocity. Initially, v = b,

therefore, the kinetic energy initially is mb2 /2. The work done deforming a spring
an amount y from its undeformed state is stored in the spring and is known as the
elastic
Rx potential
Rx energy. For our example, then, the potential energy is given by
0
F dy = 0 ky dy = kx2 /2. For x = u(0) = a, this becomes ka2 /2 as the initial
potential energy.
p p
(b) The general solution is u(t) = A cos(
p k/m t) + B sin( k/m t). The initial
conditions imply
p that A =
p a and B =
p b m/k . Therefore, the specific solution is
u(t) = a cos( k/m t) + b m/k sin( k/m t).

(c) By the equation for u, we see that

ku2 k p p p p
= (A2 cos2 ( k/m t) + 2AB cos( k/m t) sin( k/m t) + B 2 sin2 ( k/m t)).
2 2
p p p p
Further, u0 (t) = −A k/m sin( k/m t) + B k/m cos( k/m t) implies
m(u0 )2 m p p p
= (A2 (k/m) sin2 ( k/m t) − 2AB(k/m) sin( k/m t) cos( k/m t)
2 2 p
+ B 2 (k/m) cos2 ( k/m t)).
Hence the total energy satisfies
ku2 m(u0 )2 k k ka2 mb2
+ = (A2 + B 2 ) = (a2 + b2 (m/k)) = + .
2 2 2 2 2 2
Therefore, energy is conserved.
80 Chapter 3. Second Order Linear Equations

3.8
1. Consider the trigonometric identities cos(α ± β) = cos α cos β ∓ sin α sin β. Sub-
tracting these two identities we obtain cos(α − β) − cos(α + β) = 2 sin α sin β. Here,
our expression is cos 9t − cos 7t. Therefore, we let α − β = 9t and α + β = 7t. Solv-
ing this system of equations, we have α = 8t and β = −t. Therefore, we can write
cos 9t − cos 7t = 2 sin 8t sin(−t) = −2 sin 8t sin t.

5. The spring constant is k = 4/(1.5/12) = 32 lb/ft. The mass is m = 4/32 =

1/8 lb-s2 /ft. Assuming no damping, but an external force, F (t) = 2 cos 3t, the
equation describing the motion is u00 /8 + 32u = 2 cos 3t which can be rewritten as
u00 + 256u = 16 cos 3t. The initial conditions are u(0) = 2/12 = 1/6 ft. and u0 (0) =
0, where u is measured in feet and t in seconds.

7.(a) The solution of the homogeneous problem is uh (t) = c1 cos 16t + c2 sin 16t.
To find a solution of the nonhomogeneous problem, we look for a solution of the
form U (t) = A cos 3t (since there is no first derivative term, we may exclude the
sin 3t function). Looking for a solution of this form, we arrive at the equation
−9A cos 3t + 256A cos 3t = 16 cos 3t. Therefore, we need A to satisfy 247A = 16 or
A = 16/247. Therefore, the solution of the nonhomogeneous problem is u(t) =
c1 cos 16t + c2 sin 16t + 16 cos 3t/247. The initial conditions are u(0) = 1/6 and
u0 (0) = 0. Therefore, c1 , c2 must satisfy c1 + 16/247 = 1/6 and 16c2 = 0. So c1 =
151/1482 and c2 = 0. The solution is u(t) = (151/1482) cos 16t + (16/247) cos 3t.

(b)

(c) Resonance occurs when the frequency ω of the forcing function 4 sin ωt is the
same as the natural frequency ω0 of the system. Since ω0 = 16, the system will
resonate when ω = 16 rad/sec.

10. The spring constant is k = 8/(1/2) = 16 lb/ft and the mass is 8/32 = 1/4 lb-
s2 /ft. The forcing term is 8 sin 8t. Therefore, the equation of motion is u00 /4 +
16u = 8 sin 8t, which can be simplified to u00 + 64u = 32 sin 8t. The solution of the
homogeneous problem is uh (t) = c1 cos 8t + c2 sin 8t. Then we look for a particular
solution of the form up (t) = At sin 8t + Bt cos 8t. Substituting up into the ODE, we
conclude that A = 0 and B = −2. Therefore, the general solution of this differential
3.8 81

equation is u(t) = c1 cos 8t + c2 sin 8t − 2t cos 8t. The initial conditions u(0) = 1/4
ft and u0 (0) = 0 ft/sec imply that c1 = 1/4 and c2 = 1/4. Therefore, the solution
of this IVP is u(t) = (1/4) cos 8t + (1/4) sin 8t − 2t cos 8t. Solving for u0 , we see
that u0 (t) = (−2 + 16t) sin 8t. We see that u0 (t) = 0 when t = 1/8 or sin 8t = 0.
Therefore, the first four times the velocity is zero are t = 1/8, π/8, π/4, 3π/8.

11.(a) The spring constant is k = 8/(1/2) = 16 lb/ft and the mass is 8/32 = 1/4
lb-s2 /ft. The damping constant is γ = 0.25 lb-sec/ft. The external force is 4 cos 2t
lbs. Therefore, the equation of motion is u00 /4 + u0 /4 + 16u = 4 cos 2t which can
be simplified to√u00 + u0 + 64u = 16 cos 2t. The roots of the characteristic equation
are r = (−1 ± 255i)/2. Therefore, the solution of the homogeneous equation will
be transient. To find the steady-state solution, we look for a particular solution
of the form up (t) = A cos 2t + B sin 2t. Substituting up into the ODE, we conclude
that A = 240/901 and B = 8/901. Therefore, the steady-state response is u(t) =
(240/901) cos 2t + (8/901) sin 2t.

(b) With a forcing term of the form F0 cos(ωt), the steady-state response can be
written as U (t) = R cos(ωt − δ) where the amplitude

F0 F0
R= p =p .
m2 (ω02 − ω 2 )2 + γ 2 ω2 (k − mω 2 )2 + γ 2 ω 2

The amplitude will be maximized when the denominator is minimized. This will
occur when k = mω 2 ; that is, when m = k/ω 2 = 16/4 = 4 slugs.

14. Since U (t) = R cos(ωt − δ) we have U 0 (t) = −(F0 ω/∆) sin(ωt − δ), where ∆ is
given by Eq.(12). Since F0 is a constant, differentiate ω/∆ with respect to ω and
set it equal to zero. Alternatively, we can minimize (∆/ω)2 , which simplifies the
differentiation.

15. First, for 0 ≤ t ≤ π, the solution of the homogeneous problem is uh (t) =

c1 cos t + c2 sin t. Then we look for a particular solution of the form up (t) = At. We
see that a particular solution is given by up (t) = F0 t. Therefore, the general solution
for 0 ≤ t ≤ π is u(t) = c1 cos t + c2 sin t + F0 t. The initial condition u(0) = 0 im-
plies c1 = 0. Then u0 (t) = −c1 sin t + c2 cos t + F0 . Therefore, u0 (0) = c2 + F0 = 0
implies c2 = −F0 . Therefore, the solution of this IVP is u(t) = −F0 sin t + F0 t
for 0 ≤ t ≤ π. Then at time t = π, u(π) = F0 π and u0 (π) = 2F0 . In this time
interval, the forcing term is F (t) = F0 (2π − t). Therefore, we look for a partic-
ular solution of the form up (t) = B + Ct. Substituting a function of this form
into the ODE, we see that B = 2πF0 and C = −F0 . Therefore, the general so-
lution for π < t − 2π is u(t) = c1 cos t + c2 sin t + 2πF0 − F0 t. Now considering
the initial conditions u(π) = F0 π and u0 (π) = 2F0 , we need −c1 + πF0 = F0 π and
−c2 − F0 = 2F0 . Therefore, c1 = 0 and c2 = −3F0 . Therefore, the solution of
the IVP for π < t ≤ 2π is u(t) = −3F0 sin t + 2πF0 − F0 t. Then at time t = 2π,
u(2π) = 0 and u0 (2π) = −4F0 . In this time interval, the forcing term is F (t) = 0.
Therefore, the general solution is given by the solution of the homogeneous prob-
lem, u(t) = c1 cos t + c2 sin t. Considering our initial conditions, we need c1 = 0
and c2 = −4F0 . Therefore, the solution of the IVP for 2π < t is given by u(t) =
82 Chapter 3. Second Order Linear Equations

−4F0 sin t. To summarize, we conclude that the solution is given by


 −F0 sin t + F0 t 0≤t≤π
u(t) = −3F0 sin t + 2πF0 − F0 t π < t ≤ 2π
−4F0 sin t 2π < t.


16. The inductance is L = 1 henry. The resistance R = 5 × 103 ohms. The ca-
pacitance C = 0.25 × 10−6 farads. The forcing term is due to the 12−volt bat-
tery. Therefore, the equation for charge Q is Q00 + 5000Q0 + (4 × 106 )Q = 12.
The initial conditions are Q(0) = 0, Q0 (0) = 0. The solution of the homogeneous
problem is Qh (t) = c1 e−1000t + c2 e−4000t . The particular solution is of the form
Qp (t) = A, so A = 3 × 10−6 . Therefore, the general solution is given by Q(t) =
c1 e−1000t + c2 e−4000t + 3 × 10−6 . Considering our initial conditions, we conclude
that c1 = −4 × 10−6 and c2 = 10−6 . Therefore, the solution of the IVP is Q(t) =
10−6 (−4e−1000t + e−4000t + 3). At t = 0.001, 0.01, we have Q(0.001) ≈ 1.5468 ×
10−6 and Q(0.01) ≈ 2.9998 × 10−6 , respectively. From our function Q, we see that
Q(t) → 3 × 10−6 as t → ∞.

22.(a) The steady-state response is 12 sin 2t and thus the amplitude of the steady-
state response is four times the amplitude of the forcing term. This large an increase
is due to the fact that the forcing function has the same frequency as the natural
frequency, ω0 = 2, of the system. The graph also shows a phase lag of approximately
1/4 of a period. That is, the maximum of the response occurs 1/4 of a period after
the maximum of the forcing function. Both these results are substantially different
than those of either Problems 21 or 23.
3.8 83

(b) Phase plot - u 0 vs u :

24.(a) For ω = 1:

(b)

(a) ω = 0.5 (b) ω = 0.75 (c) ω = 1.25

(d) ω = 1.5 (e) ω = 1.75 (f) ω = 2

84 Chapter 3. Second Order Linear Equations

From viewing the above graphs, it appears that the system exhibits a beat near
ω = 1.5, while the pattern for ω = 1 is more irregular. However, the system exhibits
the resonance characteristic of the linear system for ω near 1, as the amplitude of
the response is the largest here.