Professional Documents
Culture Documents
The series is absolutely convergent for 2|x| < 1 or |x| < 12 . The radius of convergence is
P
R = 21 . At x = − 21 , the series ∞ n
n=1 (−1) /n converges by the alternating series test. At
P
x = 12 , the series ∞ n=1 1/n is the harmonic series which diverges. Thus, the given series
1 1
converges on [− 2 , 2 ).
100n+1 (x + 7)n+1 /(n + 1)!
2. lim = lim 100 |x + 7| = 0
n→∞ n n
100 (x + 7) /n! n→∞ n + 1
The radius of convergence is R = ∞. The series is absolutely convergent on (−∞, ∞).
1
The series is absolutely convergent for 10 |x − 5| < 1, |x − 5| < 10, or on (−5, 15). The radius
P n n n
P∞
of convergence is R = 10. At x = −5, the series ∞ n=1 (−1) (−10) /10 = n=1 1 diverges
P∞ P∞
by the nth term test. At x = 15, the series n=1 (−1) 10 /10 = n=1 (−1)n diverges by
n n n
the nth term test. Thus, the series converges on (−5, 15).
4.
(
(n + 1)!(x − 1)n+1 ∞, x 6= 1
lim n
= lim (n + 1)|x − 1| =
n→∞ n!(x − 1) n→∞ 0, x=1
398
5.1 Solutions about Ordinary Points 399
−x x2 x3 x4 x2 x4 x3 x4
6. e cos x = 1 − x + − + − ··· 1− + − ··· = 1 − x + − + ···
2 6 24 2 24 3 6
1 1 x2 5x4 61x6
7. = x2 x4 x6
=1+ + + + ···
cos x 1− + − + ··· 2 4! 6!
2 4! 6!
Since cos (π/2) = cos(−π/2) = 0, the series converges on (−π/2, π/2).
1−x 1 3 32 3
8. = − x+ − x3 + · · ·
2+x 2 4 8 16
Since the function is undefined at x = −2, the series converges on (−2, 2).
∞
X ∞
X ∞
X ∞
X
n−1 n+1 0 n−1
11. 2ncn x + 6cn x = 2 · 1 · c1 x + 2ncn x + 6cn xn+1
n=1 n=0 n=2 n=0
| {z } | {z }
k=n−1 k=n+1
∞
X ∞
X
= 2c1 + 2(k + 1)ck+1 xk + 6ck−1 xk
k=1 k=1
∞
X
= 2c1 + [2(k + 1)ck+1 + 6ck−1 ]xk
k=1
∞
X ∞
X ∞
X
12. n(n − 1)cn xn + 2 n(n − 1)cn xn−2 + 3 ncn xn
n=2 n=2 n=1
= 2 · 2 · 1c2 x0 + 2 · 3 · 2c3 x1 + 3 · 1 · c1 x1
∞
X ∞
X ∞
X
n n−2
+ n(n − 1)cn x +2 n(n − 1)cn x +3 ncn xn
n=2 n=4 n=2
| {z } | {z } | {z }
k=n k=n−2 k=n
∞
X ∞
X ∞
X
= 4c2 + (3c1 + 12c3 )x + k(k − 1)ck xk + 2 (k + 2)(k + 1)ck+2 xk + 3 kck xk
k=2 k=2 k=2
∞
X
= 4c2 + (3c1 + 12c3 )x + [(k(k − 1) + 3k) ck + 2(k + 2)(k + 1)ck+2 ] xk
k=2
∞
X
= 4c2 + (3c1 + 12c3 )x + [k(k + 2)ck + 2(k + 1)(k + 2)ck+2 ] xk
k=2
400 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
∞
X ∞
X
13. y ′ = (−1)n+1 xn−1 , y ′′ = (−1)n+1 (n − 1)xn−2
n=1 n=2
∞
X ∞
X
(x + 1)y ′′ + y ′ = (x + 1) (−1)n+1 (n − 1)xn−2 + (−1)n+1 xn−1
n=2 n=1
∞
X ∞
X ∞
X
= (−1)n+1 (n − 1)xn−1 + (−1)n+1 (n − 1)xn−2 + (−1)n+1 xn−1
n=2 n=2 n=1
∞
X ∞
X ∞
X
= −x0 + x0 + (−1)n+1 (n − 1)xn−1 + (−1)n+1 (n − 1)xn−2 + (−1)n+1 xn−1
n=2 n=3 n=2
| {z } | {z } | {z }
k=n−1 k=n−2 k=n−1
∞
X ∞
X ∞
X
k+2 k k+3 k
= (−1) kx + (−1) (k + 1)x + (−1)k+2 xk
k=1 k=1 k=1
∞ h
X i
= (−1)k+2 k − (−1)k+2 k − (−1)k+2 + (−1)k+2 xk = 0
k=1
∞ ∞
X (−1)n 2n 2n−1
X (−1)n 2n(2n − 1)
′
14. y = x , ′′
y = x2n−2
22n (n!)2 22n (n!)2
n=1 n=1
∞ ∞ ∞
X (−1)n 2n(2n − 1) X (−1)n 2n X (−1)n 2n+1
xy ′′ + y ′ + xy = x2n−1 + x2n−1 + x
22n (n!)2 22n (n!)2 22n (n!)2
n=1 n=1 n=0
| {z } | {z } | {z }
k=n k=n k=n+1
∞
X (−1)k 2k(2k − 1) (−1)k 2k (−1)k−1
= + 2k + 2k−2 x2k−1
22k (k!)2 2 (k!)2 2 [(k − 1)!]2
k=1
∞
X (−1)k (2k)2 (−1)k
= − x2k−1
22k (k!)2 22k−2 [(k − 1)!]2
k=1
∞
X
k (2k)2 − 22 k2 2k−1
= (−1) x =0
22k (k!)2
k=1
15. The singular points of (x2 − 25)y ′′ + 2xy ′ + y = 0 are −5 and 5. The distance from 0 to either
of these points is 5. The distance from 1 to the closest of these points is 4.
16. The singular points of (x2 − 2x + 10)y ′′ + xy ′ − 4y = 0 are 1 + 3i and 1 − 3i. The distance
√
from 0 to either of these points is 10 . The distance from 1 to either of these points is 3.
5.1 Solutions about Ordinary Points 401
∞
X
17. Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X ∞
X
y ′′ − xy = n(n − 1)cn xn−2 − cn xn+1 = (k + 2)(k + 1)ck+2 xk − ck−1 xk
n=2 n=0 k=0 k=1
| {z } | {z }
k=n−2 k=n+1
∞
X
= 2c2 + [(k + 2)(k + 1)ck+2 − ck−1 ]xk = 0.
k=1
Thus
c2 = 0
(k + 2)(k + 1)ck+2 − ck−1 = 0
and
1
ck+2 = ck−1 , k = 1, 2, 3, . . . .
(k + 2)(k + 1)
∞
X ∞
X ∞
X ∞
X
y ′′ + x2 y = n(n − 1)cn xn−2 + cn xn+2 = (k + 2)(k + 1)ck+2 xk + ck−2 xk
n=2 n=0 k=0 k=2
| {z } | {z }
k=n−2 k=n+2
∞
X
= 2c2 + 6c3 x + [(k + 2)(k + 1)ck+2 + ck−2 ]xk = 0.
k=2
Thus
c2 = c3 = 0
(k + 2)(k + 1)ck+2 + ck−2 = 0
402 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
and
1
ck+2 = − ck−2 , k = 2, 3, 4, . . . .
(k + 2)(k + 1)
Choosing c0 = 1 and c1 = 0 we find
1 1
c4 = − c5 = c6 = c7 = 0 c8 =
12 672
and so on. For c0 = 0 and c1 = 1 we obtain
1 1
c4 = 0 c5 = − c6 = c7 = c8 = 0 c9 =
20 1440
and so on. Thus, two solutions are
1 4 1 8 1 5 1 9
y1 = 1 − x + x − ··· and y2 = x − x + x − ··· .
12 672 20 1440
∞
X
19. Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X
n−2 n
′′ ′
y − 2xy + y = n(n − 1)cn x −2 ncn x + cn xn
n=2 n=1 n=0
| {z } | {z } | {z }
k=n−2 k=n k=n
∞
X ∞
X ∞
X
= (k + 2)(k + 1)ck+2 xk − 2 kck xk + ck xk
k=0 k=1 k=0
∞
X
= 2c2 + c0 + [(k + 2)(k + 1)ck+2 − (2k − 1)ck ]xk = 0.
k=1
Thus
2c2 + c0 = 0
(k + 2)(k + 1)ck+2 − (2k − 1)ck = 0
and
1
c2 = − c0
2
2k − 1
ck+2 = ck , k = 1, 2, 3, . . . .
(k + 2)(k + 1)
Choosing c0 = 1 and c1 = 0 we find
1 1 7
c2 = − c3 = c5 = c7 = · · · = 0 c4 = − c6 = −
2 8 240
and so on. For c0 = 0 and c1 = 1 we obtain
1 1
c2 = c4 = c6 = · · · = 0 c3 = c5 = c7 = 1112
6 24
and so on. Thus, two solutions are
1 1 7 6 1 1 1 7
y 1 = 1 − x2 − x4 − x − ··· and y 2 = x + x3 + x5 + x + ··· .
2 8 240 6 24 112
5.1 Solutions about Ordinary Points 403
∞
X
20. Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X
y ′′ − xy ′ + 2y = n(n − 1)cn xn−2 − ncn xn +2 cn xn
n=2 n=1 n=0
| {z } | {z } | {z }
k=n−2 k=n k=n
∞
X ∞
X ∞
X
= (k + 2)(k + 1)ck+2 xk − kck xk + 2 ck xk
k=0 k=1 k=0
∞
X
= 2c2 + 2c0 + [(k + 2)(k + 1)ck+2 − (k − 2)ck ]xk = 0.
k=1
Thus
2c2 + 2c0 = 0
(k + 2)(k + 1)ck+2 − (k − 2)ck = 0
and
c2 = −c0
k−2
ck+2 = ck , k = 1, 2, 3, . . . .
(k + 2)(k + 1)
c2 = −1 c3 = c5 = c7 = · · · = 0 c4 = 0 c6 = c8 = c10 = · · · = 0.
∞
X ∞
X ∞
X
k k
= (k + 2)(k + 1)ck+2 x + (k − 1)ck−1 x + ck−1 xk
k=0 k=2 k=1
∞
X
= 2c2 + (6c3 + c0 )x + [(k + 2)(k + 1)ck+2 + kck−1 ]xk = 0.
k=2
404 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
Thus
c2 = 0
6c3 + c0 = 0
(k + 2)(k + 1)ck+2 + kck−1 = 0
and
c2 = 0
1
c3 = − c0
6
k
ck+2 = − ck−1 , k = 2, 3, 4, . . . .
(k + 2)(k + 1)
∞
X
22. Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X
y ′′ + 2xy ′ + 2y = n(n − 1)cn xn−2 +2 ncn xn +2 cn xn
n=2 n=1 n=0
| {z } | {z } | {z }
k=n−2 k=n k=n
∞
X ∞
X ∞
X
k k
= (k + 2)(k + 1)ck+2 x + 2 kck x + 2 ck xk
k=0 k=1 k=0
∞
X
= 2c2 + 2c0 + [(k + 2)(k + 1)ck+2 + 2(k + 1)ck ]xk = 0.
k=1
Thus
2c2 + 2c0 = 0
(k + 2)(k + 1)ck+2 + 2(k + 1)ck = 0
5.1 Solutions about Ordinary Points 405
and
c2 = −c0
2
ck+2 = − ck , k = 1, 2, 3, . . . .
k+2
Choosing c0 = 1 and c1 = 0 we find
1 1
c2 = −1 c3 = c5 = c7 = · · · = 0 c4 = c6 = −
2 6
and so on. For c0 = 0 and c1 = 1 we obtain
2
c2 = c4 = c6 = · · · = 0 c3 = − c5 = 415 c7 = −8105
3
and so on. Thus, two solutions are
1 1 2 4 8 7
y 1 = 1 − x2 + x4 − x6 + · · · and y2 = x − x3 + x5 − x + ··· .
2 6 3 15 105
∞
X
23. Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X
(x − 1)y ′′ + y ′ = n(n − 1)cn xn−1 − n(n − 1)cn xn−2 + ncn xn−1
n=2 n=2 n=1
| {z } | {z } | {z }
k=n−1 k=n−2 k=n−1
∞
X ∞
X ∞
X
k k
= (k + 1)kck+1 x − (k + 2)(k + 1)ck+2 x + (k + 1)ck+1 xk
k=1 k=0 k=0
∞
X
= −2c2 + c1 + [(k + 1)kck+1 − (k + 2)(k + 1)ck+2 + (k + 1)ck+1 ]xk = 0.
k=1
Thus
−2c2 + c1 = 0
2
(k + 1) ck+1 − (k + 2)(k + 1)ck+2 = 0
and
1
c2 = c1
2
k+1
ck+2 = ck+1 , k = 1, 2, 3, . . . .
k+2
Choosing c0 = 1 and c1 = 0 we find c2 = c3 = c4 = · · · = 0. For c0 = 0 and c1 = 1 we obtain
1 1 1
c2 = , c3 = , c4 = ,
2 3 4
and so on. Thus, two solutions are
1 1 1
y1 = 1 and y 2 = x + x2 + x3 + x4 + · · · .
2 3 4
406 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
∞
X
24. Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X ∞
X
(x + 2)y ′′ + xy ′ − y = n(n − 1)cn xn−1 + 2n(n − 1)cn xn−2 + ncn xn − cn xn
n=2 n=2 n=1 n=0
| {z } | {z } | {z } | {z }
k=n−1 k=n−2 k=n k=n
∞
X ∞
X ∞
X ∞
X
= (k + 1)kck+1 xk + 2(k + 2)(k + 1)ck+2 xk + kck xk − ck xk
k=1 k=0 k=1 k=0
∞
X
= 4c2 − c0 + [(k + 1)kck+1 + 2(k + 2)(k + 1)ck+2 + (k − 1)ck ] xk = 0.
k=1
Thus
4c2 − c0 = 0
(k + 1)kck+1 + 2(k + 2)(k + 1)ck+2 + (k − 1)ck = 0, k = 1, 2, 3, . . .
and
1
c2 = c0
4
(k + 1)kck+1 + (k − 1)ck
ck+2 = − , k = 1, 2, 3, . . . .
2(k + 2)(k + 1)
Choosing c0 = 1 and c1 = 0 we find
1 1 1
c1 = 0, c2 = , c3 = − , c4 = 0, c5 =
4 24 480
and so on. For c0 = 0 and c1 = 1 we obtain
c2 = 0 c3 = 0 c4 = c5 = c6 = · · · = 0.
∞
X ∞
X ∞
X ∞
X
k k k
= (k + 2)(k + 1)ck+2 x − kck x − (k + 1)ck+1 x − ck xk
k=0 k=1 k=0 k=0
∞
X
= 2c2 − c1 − c0 + [(k + 2)(k + 1)ck+2 − (k + 1)ck+1 − (k + 1)ck ]xk = 0.
k=1
5.1 Solutions about Ordinary Points 407
Thus
2c2 − c1 − c0 = 0
(k + 2)(k + 1)ck+2 − (k + 1)(ck+1 + ck ) = 0
and
c1 + c0
c2 =
2
ck+1 + ck
ck+2 = , k = 1, 2, 3, . . . .
k+2
∞
X ∞
X ∞
X
x2 + 1 y ′′ − 6y = n(n − 1)cn xn + n(n − 1)cn xn−2 −6 cn xn
n=2 n=2 n=0
| {z } | {z } | {z }
k=n k=n−2 k=n
∞
X ∞
X ∞
X
= k(k − 1)ck xk + (k + 2)(k + 1)ck+2 xk − 6 ck xk
k=2 k=0 k=0
∞
X
= 2c2 − 6c0 + (6c3 − 6c1 )x + k2 − k − 6 ck + (k + 2)(k + 1)ck+2 xk = 0.
k=2
Thus
2c2 − 6c0 = 0
6c3 − 6c1 = 0
(k − 3)(k + 2)ck + (k + 2)(k + 1)ck+2 = 0
and
k−3
c2 = 3c0 c3 = c1 ck+2 = − ck , k = 2, 3, 4, . . . .
k+1
408 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
c2 = c4 = c6 = · · · = 0 c3 = 1 c5 = c7 = c9 = · · · = 0.
∞
X ∞
X ∞
X ∞
X
x2 + 2 y ′′ + 3xy ′ − y = n(n − 1)cn xn +2 n(n − 1)cn xn−2 +3 ncn xn − cn xn
n=2 n=2 n=1 n=0
| {z } | {z } | {z } | {z }
k=n k=n−2 k=n k=n
∞
X ∞
X ∞
X ∞
X
k k k
= k(k − 1)ck x + 2 (k + 2)(k + 1)ck+2 x + 3 kck x − ck xk
k=2 k=0 k=1 k=0
∞
X
= (4c2 − c0 ) + (12c3 + 2c1 )x + 2(k + 2)(k + 1)ck+2 + k2 + 2k − 1 ck xk = 0.
k=2
Thus
4c2 − c0 = 0
12c3 + 2c1 = 0
2
2(k + 2)(k + 1)ck+2 + k + 2k − 1 ck = 0
and
1 1 k2 + 2k − 1
c2 = c0 c3 = − c1 ck+2 = − ck , k = 2, 3, 4, . . . .
4 6 2(k + 2)(k + 1)
∞
X
28. Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X ∞
X
2
′′ n n−2 n
′
x − 1 y + xy − y = n(n − 1)cn x − n(n − 1)cn x + ncn x − cn xn
n=2 n=2 n=1 n=0
| {z } | {z } | {z } | {z }
k=n k=n−2 k=n k=n
∞
X ∞
X ∞
X ∞
X
= k(k − 1)ck xk − (k + 2)(k + 1)ck+2 xk + kck xk − ck xk
k=2 k=0 k=1 k=0
∞
X
= (−2c2 − c0 ) − 6c3 x + −(k + 2)(k + 1)ck+2 + k2 − 1 ck xk = 0.
k=2
Thus
−2c2 − c0 = 0
−6c3 = 0
−(k + 2)(k + 1)ck+2 + (k − 1)(k + 1)ck = 0
and
1 k−1
c2 = − c0 c3 = 0 ck+2 = ck , k = 2, 3, 4, . . . .
2 k+2
Choosing c0 = 1 and c1 = 0 we find
1 1
c2 = − c3 = c5 = c7 = · · · = 0 c4 = −
2 8
and so on. For c0 = 0 and c1 = 1 we obtain
c2 = c4 = c6 = · · · = 0 c3 = c5 = c7 = · · · = 0.
∞
X ∞
X ∞
X ∞
X
k k k
= (k + 1)kck+1 x − (k + 2)(k + 1)ck+2 x − kck x + ck xk
k=1 k=0 k=1 k=0
∞
X
= −2c2 + c0 + [−(k + 2)(k + 1)ck+2 + (k + 1)kck+1 − (k − 1)ck ]xk = 0.
k=1
410 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
Thus
−2c2 + c0 = 0
−(k + 2)(k + 1)ck+2 + (k − 1)kck+1 − (k − 1)ck = 0
and
1
c2 = c0
2
kck+1 (k − 1)ck
ck+2 = − , k = 1, 2, 3, . . . .
k+2 (k + 2)(k + 1)
and
1
′
y = C1 x + x2 + · · · + C2 .
2
∞
X
30. Substituting y = cn xn into the differential equation we have
n=0
(x + 1)y ′′ − (2 − x)y ′ + y
∞
X ∞
X ∞
X ∞
X ∞
X
= n(n − 1)cn xn−1 + n(n − 1)cn xn−2 −2 ncn xn−1 + ncn xn + cn xn
n=2 n=2 n=1 n=1 n=0
| {z } | {z } | {z } | {z } | {z }
k=n−1 k=n−2 k=n−1 k=n k=n
∞
X ∞
X ∞
X ∞
X ∞
X
= (k + 1)kck+1 xk + (k + 2)(k + 1)ck+2 xk − 2 (k + 1)ck+1 xk + kck xk + ck xk
k=1 k=0 k=0 k=1 k=0
∞
X
= 2c2 − 2c1 + c0 + [(k + 2)(k + 1)ck+2 − (k + 1)ck+1 + (k + 1)ck ]xk = 0.
k=1
Thus
and
1 1 1
c2 = c1 − c0 ck+2 = ck+1 − ck , k = 1, 2, 3, . . . .
2 k+2 k+2
∞
X ∞
X ∞
X
k k
= (k + 2)(k + 1)ck+2 x − 2 kck x + 8 ck xk
k=0 k=1 k=0
∞
X
= 2c2 + 8c0 + [(k + 2)(k + 1)ck+2 + (8 − 2k)ck ]xk = 0.
k=1
Thus
2c2 + 8c0 = 0
(k + 2)(k + 1)ck+2 + (8 − 2k)ck = 0
and
2(k − 4)
c2 = −4c0 ck+2 = ck , k = 1, 2, 3, . . . .
(k + 2)(k + 1)
412 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
∞
X ∞
X ∞
X
k k
= k(k − 1)ck x + (k + 2)(k + 1)ck+2 x + 2kck xk
k=2 k=0 k=1
∞
X
= 2c2 + (6c3 + 2c1 )x + [k(k + 1)ck + (k + 2)(k + 1)ck+2 ] xk = 0.
k=2
Thus
2c2 = 0
6c3 + 2c1 = 0
k(k + 1)ck + (k + 2)(k + 1)ck+2 = 0
and
1 k
c2 = 0 c3 = − c1 ck+2 = − ck , k = 2, 3, 4, . . . .
3 k+2
and
y ′ = c1 1 − x2 + x4 − x6 + · · · .
1 1 1
y = x − x3 + x5 − x7 + · · · .
3 5 7
∞
X
33. Substituting y = cn xn into the differential equation we have
n=0
∞
X 1 1 5
y ′′ + (sin x)y = n(n − 1)cn xn−2 + x − x3 + x − ··· c0 + c1 x + c2 x2 + · · ·
n=2
6 120
1
= 2c2 + 6c3 x + 12c4 x2 + 20c5 x3 + · · · + c0 x + c1 x2 + c2 − c0 x3 + · · ·
6
1
= 2c2 + (6c3 + c0 )x + (12c4 + c1 )x2 + 20c5 + c2 − c0 x3 + · · · = 0.
6
Thus
1
2c2 = 0 6c3 + c0 = 0 12c4 + c1 = 0 20c5 + c2 − c0 = 0
6
and
1 1 1 1
c2 = 0 c3 = − c0 c4 = − c1 c5 = − c2 + c0 .
6 12 20 120
1 1
c2 = 0, c3 = − , c4 = 0, c5 =
6 120
1
c2 = 0, c3 = 0, c4 = − , c5 = 0
12
1 1 5 1 4
y 1 = 1 − x3 + x + ··· and y2 = x − x + ··· .
6 120 12
414 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
∞
X
34. Substituting y = cn xn into the differential equation we have
n=0
∞
X
y ′′ + ex y ′ − y = n(n − 1)cn xn−2
n=2
∞
1 1 X
+ 1 + x + x2 + x3 + · · · c1 + 2c2 x + 3c3 x2 + 4c4 x3 + · · · − cn xn
2 6 n=0
= 2c2 + 6c3 x + 12c4 x2 + 20c5 x3 + · · ·
1
+ c1 + (2c2 + c1 )x + 3c3 + 2c2 + c1 x2 + · · · − [c0 + c1 x + c2 x2 + · · · ]
2
1
= (2c2 + c1 − c0 ) + (6c3 + 2c2 )x + 12c4 + 3c3 + c2 + c1 x2 + · · · = 0.
2
Thus
2c2 + c1 − c0 = 0
6c3 + 2c2 = 0
1
12c4 + 3c3 + c2 + c1 = 0
2
and
1 1 1 1 1 1
c2 = c0 − c1 c3 = − c2 c4 = − c3 + c2 − c1 .
2 2 3 4 12 24
1 1
c2 = , c3 = − , c4 = 0
2 6
1 1 1
c2 = − , c3 = , c4 = −
2 6 24
1 1 1 1 1
y 1 = 1 + x2 − x3 + · · · and y 2 = x − x2 + x3 − x4 + · · · .
2 6 2 6 24
35. The singular points of (cos x)y ′′ + y ′ + 5y = 0 are odd integer multiples of π/2. The distance
from 0 to either ±π/2 is π/2. The singular point closest to 1 is π/2. The distance from 1 to
the closest singular point is then π/2 − 1.
5.1 Solutions about Ordinary Points 415
∞
X
36. Substituting y = cn xn into the first differential equation leads to
n=0
∞
X ∞
X ∞
X ∞
X
y ′′ − xy = n(n − 1)cn xn−2 − cn xn+1 = (k + 2)(k + 1)ck+2 xk − ck−1 xk
n=2 n=0 k=0 k=1
| {z } | {z }
k=n−2 k=n+1
∞
X
= 2c2 + [(k + 2)(k + 1)ck+2 − ck−1 ]xk = 1.
k=1
Thus
2c2 = 1
(k + 2)(k + 1)ck+2 − ck−1 = 0
and
1 ck−1
c2 = ck+2 = , k = 1, 2, 3, . . . .
2 (k + 2)(k + 1)
∞
X
Substituting y = cn xn into the second differential equation leads to
n=0
∞
X ∞
X ∞
X
y ′′ − 4xy ′ − 4y = n(n − 1)cn xn−2 − 4ncn xn − 4cn xn
n=2 n=1 n=0
| {z } | {z } | {z }
k=n−2 k=n k=n
∞
X ∞
X ∞
X
= (k + 2)(k + 1)ck+2 xk − 4kck xk − 4ck xk
k=0 k=1 k=0
∞
X
= 2c2 − 4c0 + [(k + 2)(k + 1)ck+2 − 4(k + 1)ck ] xk
k=1
∞
X 1 k
= ex = 1 + x .
k!
k=1
416 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
Thus
2c2 − 4c0 = 1
1
(k + 2)(k + 1)ck+2 − 4(k + 1)ck =
k!
and
1
c2 = + 2c0
2
1 4
ck+2 = + ck , k = 1, 2, 3, . . . .
(k + 2)! k + 2
37. We identify P (x) = 0 and Q(x) = sin x/x. The Taylor series representation for sin x/x is
1 − x2 /3! + x4 /5! − · · · , for |x| < ∞. Thus, Q(x) is analytic at x = 0 and x = 0 is an ordinary
point of the differential equation.
38. In this section, Airy’s differential equation was changed to y ′′ − xy = 0. If this equation is
written as y ′′ = xy, then if x > 0 and y > 0, then y ′′ > 0. This indicates that the solution
curve is concave up. If x < 0 and y > 0, then y ′′ < 0 then the solution curve is concave down.
If x < 0 and y < 0, then y ′′ > 0 and the curve is again concave up.
5.1 Solutions about Ordinary Points 417
∞
X
39. (a) Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X
n−2 n
′′ ′
y + xy + y = n(n − 1)cn x + ncn x + cn xn
n=2 n=1 n=0
| {z } | {z } | {z }
k=n−2 k=n k=n
∞
X ∞
X ∞
X
= (k + 2)(k + 1)ck+2 xk + kck xk + ck xk
k=0 k=1 k=0
∞
X
= (2c2 + c0 ) + [(k + 2)(k + 1)ck+2 + (k + 1)ck ] xk = 0.
k=1
Thus
2c2 + c0 = 0
(k + 2)(k + 1)ck+2 + (k + 1)ck = 0
and
1 1
c2 = − c0 ck+2 = − ck , k = 1, 2, 3, . . . .
2 k+2
c2 = c4 = c6 = · · · = 0
1 2
c3 = − = −
3 3!
1 1 1 4·2
c5 = − − = =
5 3 5·3 5!
1 4·2 6·4·2
c7 = − =−
7 5! 7!
(c) y1 y2
4 4
2 2
x x
–4 –2 2 4 –4 –2 2 4
–2 –2
–4 –4
The graphs of y1 and y2 obtained from a numerical solver are shown. We see that the
partial sum representations indicate the even and odd natures of the solution, but don’t
really give a very accurate representation of the true solution. Increasing N to about 20
gives a much more accurate representation on [−4, 4].
∞
X ∞
X ∞
X
2 /2
(d) From ex = xk /k! we see that e−x = (−x2 /2)k /k! = (−1)k x2k /2k k! . From
k=0 k=0 k=0
(5) of Section 3.2 we have
Z R Z 2 Z 2 Z
e− x dx 2 /2 e−x /2 2 e−x /2 2 2 /2
y2 = y1 dx = e−x dx = e−x /2 dx = e−x /2 ex dx
y12 (e−x2 /2 )2 e−x2
∞ Z X
∞ ∞
! ∞ Z
!
X (−1)k 2k 1 X (−1)k X 1
= k
x k
x2k dx = x2k x2k dx
2 k! 2 k! 2k k! k
2 k!
k=0 k=0 k=0 k=0
∞
! ∞
!
X (−1)k 2k
X 1
= x x2k+1
2k k! (2k + 1)2k k!
k=0 k=0
5.1 Solutions about Ordinary Points 419
1 1 1 1 3 1 1
= 1 − x2 + 2 x4 − 3 x6 + · · · x+ x + 2
x5 + x7 + · · ·
2 2 ·2 2 · 3! 3·2 5·2 ·2 7 · 23 · 3!
∞
X (−1)k 2k k!
2 3 4·2 5 6·4·2 7
=x− x + x − x + ··· = x2k+1 .
3! 5! 7! (2k + 1)!
k=0
Then
1 1 1 1
30c6 + c4 + c0 − c2 = 0 and 42c7 + c5 + c1 − c3 = 0,
24 2 24 2
1 1 1
y1 (x) = 1 − x2 + x4 − x6 + · · ·
2 12 80
and
1 1 19 7
y2 (x) = x − x3 + x5 − x + ··· .
6 30 5040
(b) From part (a) the general solution of the differential equation is y = c1 y1 + c2 y2 . Then
y(0) = c1 + c2 · 0 = c1 and y ′ (0) = c1 · 0 + c2 = c2 , so the solution of the initial-value
problem is
1 1 1 1 1 19 7
y = y 1 + y 2 = 1 + x − x2 − x3 + x4 + x5 − x6 − x + ··· .
2 6 12 30 80 5040
420 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
(c) y y y
4 4 4
2 2 2
x x x
–6 –4 –2 2 4 6 –6 –4 –2 2 4 6 –6 –4 –2 2 4 6
–2 –2 –2
–4 –4 –4
y y y
4 4 4
2 2 2
x x x
–6 –4 –2 2 4 6 –6 –4 –2 2 4 6 –6 –4 –2 2 4 6
–2 –2 –2
–4 –4 –4
(d) y
6
x
–6 –4 –2 2 4 6
–2
–4
–6
In this case p(x) = 5 and q(x) = x(x − 1)2 /(x + 1). For x0 = −1 the differential equation can
be put in the form
x+1 ′
(x + 1)2 y ′′ + 5(x + 1) y + x(x + 1)y = 0.
x−1
In this case p(x) = (x + 1)/(x − 1) and q(x) = x(x + 1).
The indicial roots are 13 and −1. Since these do not differ by an integer we expect to find two
series solutions using the method of Frobenius.
14. We identify P (x) = 1/x and Q(x) = 10/x, so that p(x) = xP (x) = 1 and q(x) = x2 Q(x) =
10x. Then a0 = 1, b0 = 0, and the indicial equation is
r(r − 1) + r = r 2 = 0.
The indicial roots are 0 and 0. Since these are equal, we expect the method of Frobenius to
yield a single series solution.
422 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
∞
X
15. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
∞
X
2xy ′′ − y ′ + 2y = 2r 2 − 3r c0 xr−1 + [2(k + r − 1)(k + r)ck − (k + r)ck + 2ck−1 ] xk+r−1
k=1
= 0,
which implies
2r 2 − 3r = r(2r − 3) = 0
and
(k + r)(2k + 2r − 3)ck + 2ck−1 = 0.
The indicial roots are r = 0 and r = 3/2. For r = 0 the recurrence relation is
2ck−1
ck = − , k = 1, 2, 3, . . . ,
k(2k − 3)
and
4
c1 = 2c0 , c2 = −2c0 , c3 = c0 ,
9
and so on. For r = 3/2 the recurrence relation is
2ck−1
ck = − , k = 1, 2, 3, . . . ,
(2k + 3)k
and
2 2 4
c1 = − c0 , c2 = c0 , c3 = − c0 ,
5 35 945
and so on. The general solution on (0, ∞) is
2 4 3 3/2 2 2 2 4 3
y = C1 1 + 2x − 2x + x + · · · + C2 x 1− x+ x − x + ··· .
9 5 35 945
∞
X
16. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
2xy ′′ + 5y ′ + xy = 2r 2 + 3r c0 xr−1 + 2r 2 + 7r + 5 c1 xr
X∞
+ [2(k + r)(k + r − 1)ck + 5(k + r)ck + ck−2 ]xk+r−1
k=2
= 0,
which implies
2r 2 + 3r = r(2r + 3) = 0,
2r 2 + 7r + 5 c1 = 0,
5.2 Solutions about Singular Points 423
and
(k + r)(2k + 2r + 3)ck + ck−2 = 0.
The indicial roots are r = −3/2 and r = 0, so c1 = 0 . For r = −3/2 the recurrence relation
is
ck−2
ck = − , k = 2, 3, 4, . . . ,
(2k − 3)k
and
1 1
c2 = − c0 , c3 = 0, c4 = c0 ,
2 40
and so on. For r = 0 the recurrence relation is
ck−2
ck = − , k = 2, 3, 4, . . . ,
k(2k + 3)
and
1 1
c2 = − c0 , c3 = 0, c4 = c0 ,
14 616
and so on. The general solution on (0, ∞) is
−3/2 1 2 1 4 1 2 1 4
y = C1 x 1 − x + x + · · · + C2 1 − x + x + ··· .
2 40 14 616
∞
X
17. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
∞
1 7 X 1
4xy ′′ + y ′ + y = 4r 2 − r c0 xr−1 + 4(k + r)(k + r − 1)ck + (k + r)ck + ck−1 xk+r−1
2 2 2
k=1
= 0,
which implies
27 7
4r − r = r 4r − =0
2 2
and
1
(k + r)(8k + 8r − 7)ck + ck−1 = 0.
2
The indicial roots are r = 0 and r = 7/8. For r = 0 the recurrence relation is
2ck−1
ck = − , k = 1, 2, 3, . . . ,
k(8k − 7)
and
2 4
c1 = −2c0 , c2 = c0 , c3 = − c0 ,
9 459
424 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
which implies
2r 2 − 3r + 1 = (2r − 1)(r − 1) = 0,
2r 2 + r c1 = 0,
and
[(k + r)(2k + 2r − 3) + 1]ck + ck−2 = 0.
The indicial roots are r = 1/2 and r = 1, so c1 = 0. For r = 1/2 the recurrence relation is
ck−2
ck = − , k = 2, 3, 4, . . . ,
k(2k − 1)
and
1 1
c2 = − c0 , c3 = 0, c4 = c0 ,
6 168
and so on. For r = 1 the recurrence relation is
ck−2
ck = − , k = 2, 3, 4, . . . ,
k(2k + 1)
and
1 1
c2 = − c0 , c3 = 0, c4 = c0 ,
10 360
and so on. The general solution on (0, ∞) is
1/2 1 2 1 4 1 2 1 4
y = C1 x 1− x + x + · · · + C2 x 1 − x + x + ··· .
6 168 10 360
5.2 Solutions about Singular Points 425
∞
X
19. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
3xy ′′ + (2 − x)y ′ − y = 3r 2 − r c0 xr−1
X∞
+ [3(k + r − 1)(k + r)ck + 2(k + r)ck − (k + r)ck−1 ]xk+r−1
k=1
= 0,
which implies
3r 2 − r = r(3r − 1) = 0
and
(k + r)(3k + 3r − 1)ck − (k + r)ck−1 = 0.
The indicial roots are r = 0 and r = 1/3. For r = 0 the recurrence relation is
ck−1
ck = , k = 1, 2, 3, . . . ,
3k − 1
and
1 1 1
c1 = c0 , c2 = c0 , c3 = c0 ,
2 10 80
and so on. For r = 1/3 the recurrence relation is
ck−1
ck = , k = 1, 2, 3, . . . ,
3k
and
1 1 1
c1 = c0 , c2 = c0 , c3 = c0 ,
3 18 162
and so on. The general solution on (0, ∞) is
1 1 2 1 3 1/3 1 1 2 1 3
y = C1 1 + x + x + x + · · · + C2 x 1+ x+ x + x + ··· .
2 10 80 3 18 162
∞
X
20. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
∞
2 ′′ 2 2 2 r
X 2
x y − x− y= r −r+ c0 x + (k + r)(k + r − 1)ck + ck − ck−1 xk+r
9 9 9
k=1
= 0,
which implies
2 2 1
r2 − r + = r− r− =0
9 3 3
and
2
(k + r)(k + r − 1) + ck − ck−1 = 0.
9
426 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
The indicial roots are r = 2/3 and r = 1/3. For r = 2/3 the recurrence relation is
3ck−1
ck = , k = 1, 2, 3, . . . ,
3k2 + k
and
3 9 9
c1 = c0 , c2 = c0 , c3 = c0 ,
4 56 560
and so on. For r = 1/3 the recurrence relation is
3ck−1
ck = , k = 1, 2, 3, . . . ,
3k2 − k
and
3 9 9
c1 = c0 , c2 = c0 , c3 = c0 ,
2 20 160
and so on. The general solution on (0, ∞) is
2/3 3 9 2 9 3 1/3 3 9 2 9 3
y = C1 x 1+ x+ x + x + · · · + C2 x 1+ x+ x + x + ··· .
4 56 560 2 20 160
∞
X
21. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
∞
X
′′ ′ 2
r−1
2xy − (3 + 2x)y + y = 2r − 5r c0 x + [2(k + r)(k + r − 1)ck
k=1
− 3(k + r)ck − 2(k + r − 1)ck−1 + ck−1 ]xk+r−1
= 0,
which implies
2r 2 − 5r = r(2r − 5) = 0
and
(k + r)(2k + 2r − 5)ck − (2k + 2r − 3)ck−1 = 0.
The indicial roots are r = 0 and r = 5/2. For r = 0 the recurrence relation is
(2k − 3)ck−1
ck = , k = 1, 2, 3, . . . ,
k(2k − 5)
and
1 1 1
c1 = c0 , c2 = − c0 , c3 = − c0 ,
3 6 6
and so on. For r = 5/2 the recurrence relation is
2(k + 1)ck−1
ck = , k = 1, 2, 3, . . . ,
k(2k + 5)
5.2 Solutions about Singular Points 427
and
4 4 32
c1 = c0 , c2 = c0 , c3 = c0 ,
7 21 693
and so on. The general solution on (0, ∞) is
1 1 1 4 4 32 3
y = C1 1 + x − x2 − x3 + · · · + C2 x5/2 1 + x + x2 + x + ··· .
3 6 6 7 21 693
∞
X
22. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
2 ′′ 4 2 2 4 r 2 5
′
x y + xy + x − y= r − c0 x + r + 2r + c1 xr+1
9 9 9
∞
X 4
+ (k + r)(k + r − 1)ck + (k + r)ck − ck + ck−2 xk+r
9
k=2
= 0,
which implies
2 4 2 2
r − = r+ r− = 0,
9 3 3
2 5
r + 2r + c1 = 0,
9
and
2 4
(k + r) − ck + ck−2 = 0.
9
The indicial roots are r = −2/3 and r = 2/3, so c1 = 0. For r = −2/3 the recurrence relation
is
9ck−2
ck = − , k = 2, 3, 4, . . . ,
3k(3k − 4)
and
3 9
c2 = − c0 , c3 = 0, c4 = c0 ,
4 128
and so on. For r = 2/3 the recurrence relation is
9ck−2
ck = − , k = 2, 3, 4, . . . ,
3k(3k + 4)
and
3 9
c2 = − c0 , c3 = 0, c4 = c0 ,
20 1,280
and so on. The general solution on (0, ∞) is
−2/3 3 2 9 4 2/3 3 2 9 4
y = C1 x 1− x + x + · · · + C2 x 1− x + x + ··· .
4 128 20 1,280
428 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
∞
X
23. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
9x2 y ′′ + 9x2 y ′ + 2y = 9r 2 − 9r + 2 c0 xr
X∞
+ [9(k + r)(k + r − 1)ck + 2ck + 9(k + r − 1)ck−1 ]xk+r
k=1
= 0,
which implies
9r 2 − 9r + 2 = (3r − 1)(3r − 2) = 0
and
[9(k + r)(k + r − 1) + 2]ck + 9(k + r − 1)ck−1 = 0.
The indicial roots are r = 1/3 and r = 2/3. For r = 1/3 the recurrence relation is
(3k − 2)ck−1
ck = − , k = 1, 2, 3, . . . ,
k(3k − 1)
and
1 1 7
c1 = − c0 , c2 = c0 , c3 = − c0 ,
2 5 120
and so on. For r = 2/3 the recurrence relation is
(3k − 1)ck−1
ck = − , k = 1, 2, 3, . . . ,
k(3k + 1)
and
1 5 1
c1 = − c0 , c2 = c0 , c3 = − c0 ,
2 28 21
and so on. The general solution on (0, ∞) is
1/3 1 1 2 7 3 2/3 1 5 2 1 3
y = C1 x 1− x+ x − x + · · · + C2 x 1 − x + x − x + ··· .
2 5 120 2 28 21
∞
X
24. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
2x2 y ′′ + 3xy ′ + (2x − 1)y = 2r 2 + r − 1 c0 xr
X∞
+ [2(k + r)(k + r − 1)ck + 3(k + r)ck − ck + 2ck−1 ]xk+r
k=1
= 0,
which implies
2r 2 + r − 1 = (2r − 1)(r + 1) = 0
5.2 Solutions about Singular Points 429
and
[(k + r)(2k + 2r + 1) − 1]ck + 2ck−1 = 0.
The indicial roots are r = −1 and r = 1/2. For r = −1 the recurrence relation is
2ck−1
ck = − , k = 1, 2, 3, . . . ,
k(2k − 3)
and
4
c1 = 2c0 , c2 = −2c0 , c3 = c0 ,
9
and so on. For r = 1/2 the recurrence relation is
2ck−1
ck = − , k = 1, 2, 3, . . . ,
k(2k + 3)
and
2 2 4
c1 = − c0 , c2 = c0 , c3 = − c0 ,
5 35 945
and so on. The general solution on (0, ∞) is
−1 2 4 3 1/2 2 2 2 4 3
y = C1 x 1 + 2x − 2x + x + · · · + C2 x 1− x+ x − x + ··· .
9 5 35 945
∞
X
25. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
xy ′′ + 2y ′ − xy = r 2 + r c0 xr−1 + r 2 + 3r + 2 c1 xr
X∞
+ [(k + r)(k + r − 1)ck + 2(k + r)ck − ck−2 ]xk+r−1
k=2
= 0,
which implies
r 2 + r = r(r + 1) = 0,
r 2 + 3r + 2 c1 = 0,
and
(k + r)(k + r + 1)ck − ck−2 = 0.
The indicial roots are r1 = 0 and r2 = −1, so c1 = 0. For r1 = 0 the recurrence relation is
ck−2
ck = , k = 2, 3, 4, . . . ,
k(k + 1)
and
1 1 1
c2 = c0 c3 = c5 = c7 = · · · = 0 c4 = c0 c2n = c0 .
3! 5! (2n + 1)!
430 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
and
1 1 1
c2 = c0 c3 = c5 = c7 = · · · = 0 c4 = c0 c2n = c0 .
2! 4! (2n)!
1
= [C1 sinh x + C2 cosh x] .
x
∞
X
26. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
2 ′′ 1 2 2 1 r 3
2
′
x y + xy + x − y= r − c0 x + r + 2r + c1 xr+1
4 4 4
∞
X 1
+ (k + r)(k + r − 1)ck + (k + r)ck − ck + ck−2 xk+r
4
k=2
= 0,
which implies
1
2 1 1 2 3
r − = r− r+ = 0, r + 2r + c1 = 0,
4 2 2 4
and
12
(k + r) − ck + ck−2 = 0.
4
The indicial roots are r1 = 1/2 and r2 = −1/2, so c1 = 0. For r1 = 1/2 the recurrence
relation is
ck−2
ck = − , k = 2, 3, 4, . . . ,
k(k + 1)
and
1 1 (−1)n
c2 = − c0 c3 = c5 = c7 = · · · = 0 c4 = c0 c2n = c0 .
3! 5! (2n + 1)!
5.2 Solutions about Singular Points 431
∞
X
27. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
∞
X
xy ′′ − xy ′ + y = r 2 − r c0 xr−1 + [(k + r + 1)(k + r)ck+1 − (k + r)ck + ck ]xk+r = 0
k=0
which implies
r 2 − r = r(r − 1) = 0
and
(k + r + 1)(k + r)ck+1 − (k + r − 1)ck = 0.
y = C1 x + C2 y2 (x).
432 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
∞
X
28. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
3 ′
y ′′ + y − 2y = r 2 + 2r c0 xr−2 + r 2 + 4r + 3 c1 xr−1
x
∞
X
+ [(k + r)(k + r − 1)ck + 3(k + r)ck − 2ck−2 ] xk+r−2
k=2
= 0,
which implies
r 2 + 2r = r(r + 2) = 0
r 2 + 4r + 3 c1 = 0
(k + r)(k + r + 2)ck − 2ck−2 = 0.
The indicial roots are r1 = 0 and r2 = −2, so c1 = 0. For r1 = 0 the recurrence relation is
2ck−2
ck = , k = 2, 3, 4, . . . ,
k(k + 2)
and
1 1 1
c2 = c0 c3 = c5 = c7 = · · · = 0 c4 = c0 c6 = c0 .
4 48 1,152
The result is
1 1 1
y1 = c0 1 + x2 + x4 + x6 + · · · .
4 48 1,152
A second solution is
Z − R (3/x)dx Z
e dx
y2 = y1 2 dx = y1 2
y1 x3 1 + 14 x2 + 48 1 4
x + ···
Z Z
dx 1 1 2 7 4 19 6
= y1 = y1 1− x + x + x + · · · dx
x3 1 + 12 x2 + 485 4
x + 5767 6
x + ··· x3 2 48 576
Z
1 1 7 19 3 1 1 7 2 19 4
= y1 − + x − x + · · · dx = y 1 − − ln x + x − x + · · ·
x3 2x 48 576 2x2 2 96 2,304
1 1 7 19 4
= − y1 ln x + y − 2 + x2 − x + ··· .
2 2x 96 2,304
y = C1 y1 (x) + C2 y2 (x).
5.2 Solutions about Singular Points 433
P∞
29. Substituting y = n=0 cn xn+r into the differential equation and collecting terms, we obtain
∞
X
2 r−1
′′ ′
xy + (1 − x)y − y = r c0 x + [(k + r)(k + r − 1)ck + (k + r)ck − (k + r)ck−1 ]xk+r−1 = 0,
k=1
One solution is
1 1
y1 = c0 1 + x + x2 + x3 + · · · = c0 ex .
2 3!
A second solution is
Z − R (1/x−1) dx Z x Z
e x e /x x 1 −x
y2 = y1 dx = e dx = e e dx
e2x e2x x
Z Z
1 1 1 1 1 1
= ex 1 − x + x2 − x3 + · · · dx = ex − 1 + x − x2 + · · · dx
x 2 3! x 2 3!
∞
x 1 2 1 3 X (−1)n+1 n
= e ln x − x + x − x + · · · = ex ln x − ex x .
2·2 3 · 3! n · n!
n=1
∞
X
30. Substituting y = cn xn+r into the differential equation and collecting terms, we obtain
n=0
∞
X
2 r−1
′′ ′
xy + y + y = r c0 x + [(k + r)(k + r − 1)ck + (k + r)ck + ck−1 ]xk+r−1 = 0
k=1
One solution is
∞
1 1 3 1 4 X (−1)n
y1 = c0 1 − x + 2 x2 − 2
x + x − ··· = c0 xn .
2 (3!) (4!)2 n=0
(n!)2
434 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
A second solution is
Z R Z
e− (1/x) dx dx
y2 = y1 dx = y1 2
y12 x 1−x+ 1 2 1 3
4x − 36 x + ···
Z
dx
= y1 3 2
x 1 − 2x + 35 4
2x − 95 x3 +
288 x − · · ·
Z
1 5 2 23 3 677 4
= y1 1 + 2x + x + x + x + · · · dx
x 2 9 288
Z
1 5 23 2 677 3
= y1 +2+ x+ x + x + · · · dx
x 2 9 288
5 2 23 3 677 4
= y1 ln x + 2x + x + x + x + ···
4 27 1,152
y = C1 y1 (x) + C2 y2 (x).
P∞ n+r
31. Substituting y = n=0 cn x into the differential equation and collecting terms, we obtain
∞
X
′′ ′ 2 r−1
xy + (x − 6)y − 3y = (r − 7r)c0 x + [(k + r)(k + r − 1)ck + (k + r − 1)ck−1
k=1
which implies
r 2 − 7r = r(r − 7) = 0
and
(k + r)(k + r − 7)ck + (k + r − 4)ck−1 = 0.
(k + 7)kck + (k + 3)ck−1 = 0, k = 1, 2, 3, . . . ,
or
k+3
ck = − ck−1 , k = 1, 2, 3, . . . .
k(k + 7)
Taking c0 6= 0 we obtain
1 5 1
c1 = − c0 c2 = c0 c3 = − c0 ,
2 18 6
and so on. Thus, the indicial root r1 = 7 yields a single solution. Now, for r2 = 0 the
recurrence relation is
Then
and
and
k−4
ck = − ck−1 , k = 8, 9, 10, . . . .
k(k − 7)
which implies
4r − r 2 = r(4 − r) = 0
and
−(k + r)(k + r − 4)ck + [(k + r − 1)(k + r − 2) − 2]ck−1 = 0.
or
k+1
ck = ck−1 , k = 1, 2, 3, . . . .
k
436 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
Taking c0 6= 0 we obtain
and so on. Thus, the indicial root r1 = 4 yields a single solution. For r2 = 0 the recurrence
relation is
−k(k − 4)ck + k(k − 3)ck−1 = 0, k = 1, 2, 3, . . . ,
or
−(k − 4)ck + (k − 3)ck−1 = 0, k = 1, 2, 3, . . . .
Then
and
(k − 3)ck−1
ck = , k = 5, 6, 7, . . . .
k−4
Taking c0 6= 0 and c4 = 0 we obtain
2 1
c1 = c0 c2 = c0 c3 = c4 = c5 = · · · = 0.
3 3
dy dy dt dy
= = −t2
dx dt dx dt
and
d2 y d dy d 2dy d2 y dt
2 dy dt d2 y dy
= = −t = −t − 2t = t4 + 2t3 .
dx2 dx dx dx dt dt2 dx dt dx dt2 dt
Now
4d2 y 1 2
4 d y 3 dy d2 y 2 dy
x 2
+ λy = 4 t + 2t + λy = + + λy = 0
dx t dt2 dt dt2 t dt
becomes
d2 y dy
t +2 + λty = 0.
dt2 dt
5.2 Solutions about Singular Points 437
∞
X
(b) Substituting y = cn tn+r into the differential equation and collecting terms, we obtain
n=0
d2 y dy
t +2 + λty = (r 2 + r)c0 tr−1 + (r 2 + 3r + 2)c1 tr
dt2 dt
∞
X
+ [(k + r)(k + r − 1)ck + 2(k + r)ck + λck−2 ]tk+r−1
k=2
= 0,
which implies
r 2 + r = r(r + 1) = 0,
r 2 + 3r + 2 c1 = 0,
and
(k + r)(k + r + 1)ck + λck−2 = 0.
The indicial roots are r1 = 0 and r2 = −1, so c1 = 0. For r1 = 0 the recurrence relation
is
λck−2
ck = − , k = 2, 3, 4, . . . ,
k(k + 1)
and
λ λ2 λn
c2 = − c0 c3 = c5 = c7 = · · · = 0 c4 = c0 c2n = (−1)n c0 .
3! 5! (2n + 1)!
For r2 = −1 the recurrence relation is
λck−2
ck = − , k = 2, 3, 4, . . . ,
k(k − 1)
and
λ λ2 λn
c2 = − c0 c3 = c5 = c7 = · · · = 0 c4 = c0 c2n = (−1)n c0 .
2! 4! (2n)!
1h √ √ i
= C1 sin λ t + C2 cos λ t .
t
(b) Using t = 1/x, the solution of the original equation is
√ √
λ λ
y(x) = C1 x sin + C2 x cos .
x x
438 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
or, if
n2 π 2 a2 b2 Pn b4
λn = = .
L2 EI
The critical loads are then Pn = n2 π 2 (a/b)2 EI0 /L2 . Using
√ √
C2 = −C1 sin ( λ/a)/ cos ( λ/a) we have
" √ √ √ #
λ sin ( λ/a) λ
y = C1 x sin − √ cos
x cos ( λ/a) x
" √ √ √ √ #
λ λ λ λ
= C3 x sin cos − cos sin
x a x a
√ 1 1
= C3 x sin λ − ,
x a
and
nπab 1 1 nπab a nπb a
yn (x) = C3 x sin − = C3 x sin − 1 = C4 x sin 1− .
L x a La x L x
Suppose x0 is a singular point of the differential equation. Then we say that x0 is a regular
singular point if (x − x0 )P (x), (x − x0 )2 Q(x), and (x − x0 )3 R(x) are analytic at x = x0 .
∞
X
36. Substituting y = cn xn+r into the first differential equation and collecting terms, we obtain
n=0
∞
X
3 ′′ r
x y + y = c0 x + [ck + (k + r − 1)(k + r − 2)ck−1 ] xk+r = 0.
k=1
which implies
−rc0 = 0
2
(k + r + 1) ck − (k + r + 1)ck+1 = 0.
Letting c0 = 1 we get c1 = 2, c2 = 3!, c3 = 4!, and so on. The solution of the differential
∞
X
equation is then y = (n + 1)!xn , which converges only at x = 0.
n=0
37. We write the differential equation in the form x2 y ′′ + (b/a)xy ′ + (c/a)y = 0 and identify
a0 = b/a and b0 = c/a as in (14) in the text. Then the indicial equation is
b c
r(r − 1) + r+ =0 or ar 2 + (b − a)r + c = 0,
a a
1
8. We identify α = 6 and ν = 2 . Then the general solution is y = c1 J1/2 (6x) + c2 J−1/2 (6x).
9. We identify α = 4 and ν = 2/3 . Then the general solution is y = c1 I2/3 (4x) + c2 K2/3 (4x)
√ √ √
10. We identify α = 2 and ν = 8. Then the general solution is y = c1 I8 ( 2 x) + c2 K8 ( 2 x).
1
y ′ = x−1/2 v ′ (x) − x−3/2 v(x),
2
3
y ′′ = x−1/2 v ′′ (x) − x−3/2 v ′ (x) + x−5/2 v(x),
4
and
2 ′′ ′ 2 2 3/2 ′′ 1/2 ′ 2 3/2 1 −1/2
x y + 2xy + α x y = x v (x) + x v (x) + α x − x v(x) = 0.
4
√
12. If y = x v(x) then
1
y ′ = x1/2 v ′ (x) + x−1/2 v(x)
2
1
y ′′ = x1/2 v ′′ (x) + x−1/2 v ′ (x) − x−3/2 v(x)
4
and
2 ′′ 1
2 2 2 5/2 ′′ 3/2 ′ 1 1/2 2 2 2 1
x y + α x −ν + y=x v (x) + x v (x) − x v(x) + α x − ν + x1/2 v(x)
4 4 4
13. Write the differential equation in the form y ′′ + (2/x)y ′ + (4/x)y = 0. This is the form of (18)
in the text with a = − 12 , c = 21 , b = 4, and p = 1, so, by (19) in the text, the general solution
is
y = x−1/2 [c1 J1 (4x1/2 ) + c2 Y1 (4x1/2 )].
14. Write the differential equation in the form y ′′ + (3/x)y ′ + y = 0. This is the form of (18) in
the text with a = −1, c = 1, b = 1, and p = 1, so, by (19) in the text, the general solution is
15. Write the differential equation in the form y ′′ − (1/x)y ′ + y = 0. This is the form of (18) in
the text with a = 1, c = 1, b = 1, and p = 1, so, by (19) in the text, the general solution is
16. Write the differential equation in the form y ′′ − (5/x)y ′ + y = 0. This is the form of (18) in
the text with a = 3, c = 1, b = 1, and p = 2, so, by (19) in the text, the general solution is
17. Write the differential equation in the form y ′′ + (1 − 2/x2 )y = 0. This is the form of (18) in
the text with a = 12 , c = 1, b = 1, and p = 32 , so, by (19) in the text, the general solution is
y = x1/2 [c1 J3/2 (x) + c2 Y3/2 (x)] = x1/2 [C1 J3/2 (x) + C2 J−3/2 (x)].
18. Write the differential equation in the form y ′′ + (4 + 1/4x2 )y = 0. This is the form of (18) in
the text with a = 21 , c = 1, b = 2, and p = 0, so, by (19) in the text, the general solution is
19. Write the differential equation in the form y ′′ + (3/x)y ′ + x2 y = 0. This is the form of (18) in
the text with a = −1, c = 2, b = 21 , and p = 21 , so, by (19) in the text, the general solution is
1 2 1 2
y = x−1 c1 J1/2 x + c2 Y1/2 x
2 2
or
−1 1 2 1 2
y=x C1 J1/2 x + C2 J−1/2 x .
2 2
20. Write the differential equation in the form y ′′ + (1/x)y ′ + ( 19 x4 − 4/x2 )y = 0. This is the form
of (18) in the text with a = 0, c = 3, b = 19 , and p = 32 , so, by (19) in the text, the general
solution is
1 3 1 3
y = c1 J2/3 x + c2 Y2/3 x
9 9
or
1 3 1 3
y = C1 J2/3 x + C2 J−2/3 x .
9 9
21. Using the fact that i2 = −1, along with the definition of Jν (x) in (7) in the text, we have
∞ 2n+ν
−ν −ν
X (−1)n ix
Iν (x) = i Jν (ix) = i
n!Γ(1 + ν + n) 2
n=0
∞
X (−1)n x 2n+ν
= i2n+ν−ν
n=0
n!Γ(1 + ν + n) 2
∞
X (−1)n x 2n+ν
= (i2 )n
n!Γ(1 + ν + n) 2
n=0
∞
X (−1)2n x 2n+ν
=
n!Γ(1 + ν + n) 2
n=0
∞ x 2n+ν
X 1
= ,
n=0
n!Γ(1 + ν + n) 2
22. (a) The differential equation has the form of (18) in the text with
1 − 2a = 0 2c − 2 = 2 b2 c2 = −β 2 c2 = −1 a2 − p2 c2 = 0
1 1 1 1
a= c=2 β= , b= i p=
2 2 2 4
Then, by (19) in the text,
1/2 1 2 1 2
y=x c1 J1/4 ix + c2 J−1/4 ix .
2 2
In terms of real functions the general solution can be written
1/2 1 2 1 2
y=x C1 I1/4 x + C2 K1/4 x .
2 2
5.3 Special Functions 443
(b) Write the differential equation in the form y ′′ + (1/x)y ′ − 7x2 y = 0. This is the form of
(18) in the text with
1 − 2a = 1 2c − 2 = 2 b2 c2 = −β 2 c2 = −7 a2 − p2 c2 = 0
1√ 1√
a=0 c=2 β= 7, b= 7i p = 0.
2 2
Then, by (19) in the text,
1√ 2 1√ 2
y = c1 J0 7 ix + c2 Y0 7 ix .
2 2
In terms of real functions the general solution can be written
1√ 2 1√ 2
y = C1 I0 7 x + C2 K0 7x .
2 2
23. The differential equation has the form of (18) in the text with
1 − 2a = 0 2c − 2 = 0 b2 c2 = 1 a2 − p2 c2 = 0
1 1
a= c=1 b=1 p= .
2 2
Then, by (19) in the text,
" r r #
2 2
y = x1/2 [c1 J1/2 (x) + c2 J−1/2 (x)] = x1/2 c1 sin x + c2 cos x = C1 sin x + C2 cos x.
πx πx
24. Write the differential equation in the form y ′′ + (4/x)y ′ + (1 + 2/x2 )y = 0. This is the form
of (18) in the text with
1 − 2a = 4 2c − 2 = 0 b2 c2 = 1 a2 − p2 c2 = 2
3 1
a=− c=1 b=1 p= .
2 2
Then, by (19), (23), and (24) in the text,
" r r #
−3/2 −3/2 2 2
y=x [c1 J1/2 (x) + c2 J−1/2 (x)] = x c1 sin x + c2 cos x
πx πx
1 1
= C1 2
sin x + C2 2 cos x.
x x
1 2
25. Write the differential equation in the form y ′′ + (2/x)y ′ + ( 16 x − 3/4x2 )y = 0. This is the
form of (18) in the text with
1 3
1 − 2a = 2 2c − 2 = 2 b2 c2 = a2 − p2 c2 = −
16 4
1 1 1
a=− c=2 b= p= .
2 8 2
444 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
26. Write the differential equation in the form y ′′ − (1/x)y ′ + (4 + 3/4x2 )y = 0. This is the form
of (18) in the text with
3
1 − 2a = −1 2c − 2 = 0 b2 c2 = 4 a2 − p2 c2 =
4
1
a=1 c=1 b=2 p= .
2
Then, by (19) in the text,
" r r #
2 2
y = x c1 J1/2 (2x) + c2 J−1/2 (2x) = x c1 sin 2x + c2 cos 2x
π2x π2x
(b) The formula in part (a) is a linear first-order differential equation in Jν (x). An integrating
factor for this equation is xν , so
d ν
[x Jν (x)] = xν Jν−1 (x).
dx
28. Subtracting the formula in part (a) of Problem 27 from the formula in Example 5 we obtain
0 = 2νJν (x) − xJν+1 (x) − xJν−1 (x) or 2νJν (x) = xJν+1 (x) + xJν−1 (x).
30. From (20) we obtain J0′ (x) = −J1 (x), and from (21) we obtain J0′ (x) = J−1 (x). Thus
J0′ (x) = J−1 (x) = −J1 (x).
1 (2n + 1)! √
31. (a) Using Γ 1 + + n = π we get the following
2 22n+1 n!
1 (2n + 1)! √
Γ 1+ +n = π
2 22n+1 n!
1 1 (2n + 1)! √
+n Γ +n = π
2 2 22n+1 n!
1 1 (2n + 1)! √
1− +n Γ 1− +n = π
2 2 22n+1 n!
1 (2n + 1)! √
Γ 1− +n = 1
2n+1
π
2 1− +n 2
2 n!
1 (2n + 1)! √
Γ 1− +n = 1 2n+1 n!
π
2 2 (1 + 2n) 2
1 (2n)! √
Γ 1− +n = π
2 22n n!
The last series is the Maclaurin series for the cosine therefore
r ∞ r
2 X (−1)n 2n 2
J−1/2 (x) = · x = cos x
πx (2n)! πx
n=0
(c)
32. (a) From Problem 28, 2νJν (x) = xJν+1 (x) + xJν−1 (x) and so with ν = 1/2 we get
(b)
33. Letting
r
2 k −αt/2
s= e ,
α m
we have " r # r !
dx dx ds dx 2 k α −αt/2 dx k −αt/2
= = − e = − e
dt ds dt dt α m 2 ds m
and
r ! r !
d2 x d dx dx α k −αt/2 d dx k −αt/2
2
= = e + − e
dt dt dt ds 2 m dt ds m
r ! r !
dx α k −αt/2 d2 x ds k −αt/2
= e + 2 − e
ds 2 m ds dt m
r !
dx α k −αt/2 d2 x k −αt
= e + 2 e .
ds 2 m ds m
Then r
d2 x d2 x mα k −αt/2 dx
m 2 + ke−αt x = ke−αt 2 + e + ke−αt x = 0.
dt ds 2 m ds
d2 x dx
s2 2
+s + s2 x = 0.
ds ds
448 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
2
34. (a) Differentiating y = x1/2 w 3 αx
3/2 with respect to 23 αx3/2 we obtain
2 3/2 1 2 3/2
y ′ = x1/2 w′ αx αx1/2 + x−1/2 w αx
3 2 3
and
′′ ′′ 2 3/2 1/2 ′ 2 3/2
y = αxw αx αx + αw αx
3 3
1 2 3/2 1 2 3/2
+ αw′ αx − x−3/2 w αx .
2 3 4 3
3 1
y ′′ + α2 xy = α αx3/2 w′′ + w′ + αx3/2 − w = 0.
2 4αx3/2
3 α 2 ′′ ′ 2 1
t w (t) + tw (t) + t − w(t) = 0, t > 0.
2 t 9
2
(b) Using the same substitution y = x1/2 w 3 αx
3/2 the differential equation is now y ′′ −
α2 xy = 0:
′ 1/2 2 3/2
′ 1/2 1 −1/2 2 3/2
y =x w αx αx + x w αx
3 2 3
′ 2 3/2 1 −1/2 2 3/2
= αxw αx + x w αx
3 2 3
′′ ′′ 2 3/2 1/2 ′ 2 3/2 1 ′ 2 3/2 1 −3/2 2 3/2
y = αxw αx αx + αw αx + αw αx − x w αx
3 3 2 3 4 3
2 3/2 ′′ 2 3/2 3 ′ 2 3/2 1 −3/2 2 3/2
=α x w αx + αw αx − x w αx
3 2 3 4 3
5.3 Special Functions 449
Then
′′ 2 2 3/2 ′′ 2 3/2 3 2 3/2 1 2 3/2
y − α xy = α x w αx + αw′ αx − x−3/2 w αx
3 2 3 4 3
2 3/2 2 3/2
−α x w αx
3
2 3/2 3 ′
′′ 2 3/2 1 −3/2
= α x w + αw − α x + x w
2 4
2 3/2 ′′ 3 ′ 2 3/2 1
= α x w + αw − α x + 3/2 w
2 4x
3/2 ′′ 3 ′ 3/2 1
= α αx w + w = αx + w
2 4αx3/2
3 3 3 1 2
= α tw′′ + w′ − t + t−1 w ←− t = αx3/2
2 2 2 6 3
3α 2 ′′ 1
= t w + tw′ − t2 + w =0
2t 9
when w is a solution of t2 w′′ + tw′ − t2 + 19 w = 0.
35. (a) By part (a) of Problem 34, a solution of Airy’s equation is y = x1/2 w( 23 αx3/2 ), where
is a solution of Bessel’s equation of order 13 . Thus, the general solution of Airy’s equation
for x > 0 is
1/2 2 3/2 1/2 2 3/2 1/2 2 3/2
y=x w αx = c1 x J1/3 αx + c2 x J−1/3 αx .
3 3 3
(b) By part (b) of Problem 34, a solution of Airy’s equation is y = x1/2 w( 32 αx3/2 ), where
In order to satisfy the conditions that lim y(x) and lim y ′ (x) are finite we are forced to
x→0+ x→0+
define c2 = 0. Thus, y(x) = c1 J0 (αx). The second boundary condition, y(2) = 0, implies
450 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
37. (a) The differential equation y ′′ + (λ/x)y = 0 has the form of (18) in the text with
1 − 2a = 0 2c − 2 = −1 b2 c2 = λ a2 − p2 c2 = 0
1 1 √
a= c= b=2 λ p = 1.
2 2
(b) We first note that y = J1 (t) is a solution of Bessel’s equation, t2 y ′′ + ty ′ + (t2 − 1)y = 0,
with ν = 1. That is,
t2 J1′′ (t) + tJ1′ (t) + (t2 − 1)J1 (t) = 0,
√
or, letting t = 2 x ,
√ √ √ √
4xJ1′′ 2 x + 2 xJ1′ 2 x + (4x − 1)J1 2 x = 0.
√ √
Now, if y = xJ1 (2 x ), we have
√ √ 1 1 √ √ 1 √
y′ = x J1′ 2 x √ + √ J1 2 x = J1′ 2 x + x−1/2 J1 2 x
x 2 x 2
and
√ 1 ′ √ 1 −3/2 √
y ′′ = x−1/2 J1′′ 2 x + J1 2 x − x J1 2 x .
2x 4
Then
√ √ 1 √ 1 √ √ √
xy ′′ + y = x J1′′ 2 x + J1′ 2 x − x−1/2 J1 2 x + x J 2 x
2 4
1 √ √ √ √ √
= √ 4xJ1′′ 2 x + 2 x J1′ 2 x − J1 2 x + 4xJ 2 x
4 x
= 0,
√ √
and y = x J1 (2 x ) is a solution of Airy’s differential equation.
5.3 Special Functions 451
∞ 2n+ 1
−1/2 −1/2
X (−1)n ix 2
i J1/2 (ix) = i
n=0
n!Γ 1 + + 12 + n 2
r ∞
1 X 1
= x2n+1
πx (2n + 1)
n=0
r
2
= sinh x
πx
∞ 2n− 1
1/2 1/2
X (−1)n ix 2
i J−1/2 (ix) = i 1
n!Γ 1 − 2 + n 2
n=0
r ∞
2 X 1
= x2n
πx n=0 (2n)!
r
2
= cosh x
πx
(c) Equation (16) in the text and parts (a) and (b) imply
(b)
Therefore we have
r r
π π cos x sin x
y1 (x) = Y (x) = · J−3/2 (x) = − 2 −
2x 3/2 2x x x
r r
π π 3 cos x 3 sin x cos x
y2 (x) = Y5/2 (x) = − · J−5/2 (x) = − − +
2x 2x x3 x2 x
r r
π π 15 cos x 15 sin x 6 cos x sin x
y3 (x) = J7/2 (x) = · J−7/2 (x) = − − + +
2x 2x x4 x3 x2 x
5.3 Special Functions 453
(b)
41. Using R(x) = (αx)−1/2 Z(x) and straightforward use of the product rule we get
Force these three expressions into the equation x2 R′′ (x) + 2xR′ (x) + [α2 x2 − n(n + 1)]R = 0
to get
2 −1/2 3 −5/2 −3/2 ′ −1/2 ′′
x ·α x Z(x) − x Z (x) + x Z (x)
4
1 −3/2
2x · α−1/2
x −1/2 ′
Z (x) − x Z(x) + α2 − n (n + 1) R = 0
2
1
x2 Z ′′ (x) + xZ ′ (x) + [α2 x2 − (n + )2 ]Z = 0.
2
42. (a) Compare x2 Z ′′ (x) + xZ ′ (x) + [α2 x2 − (n + 12 )2 ]Z = 0 to the parametric Bessel equation
of order ν x2 y ′′ + xy ′ + (α2 x2 − ν 2 )y = 0 whose solution is y = C1 Jν (αx) + C2 Yν (αx).
If we take ν = n + 12 then
= c1 jn (αx) + c2 yn (αx)
d2 T dT
r + − a2 r (T − Tm ) = 0
dr 2 dr
d2 T dT
r2 2
+r − a2 r 2 (T − Tm ) = 0
dr dr
If we change the dependent variable by means of the substitution w(r) = T (r) − 70 the
differential equation becomes
d2 w dw
r2 2
+r − a2 r 2 w = 0.
dr dr
This differential equation is the parametric form of the modified Bessel’s equation of order
ν = 0 so:
d d
The derivatives I0 (x) = I1 (x) and K0 (x) = −K1 (x) and so the Chain rule gives
dx dx
u
d z}|{ dI0 du d d
I0 (ar) = = I0′ (ar) ar = aI1 (ar) and K0 (ar) = −aK1 (ar)
dr du dr dr dr
Therefore
d d
T ′ (r) = c1 I0 (ar) + c2 K0 (ar) or T ′ (r) = c1 aI1 (ar) − c2 aK1 (ar)
dr dr
5.3 Special Functions 455
The boundary conditions T (1) = 160 and T ′ (3) = 0 then give the linear system of equations
c1 I0 (a) + c2 K0 (a) = 90
c1 I1 (3a) − c2 K1 (3a) = 0.
44. The differential equation is the same as in Problem 43 so the substitution w(r) = T (r) − 70
again yields
d2 w dw
r2 2 + r − a2 r 2 w = 0
dr dr
and so
The boundary conditions T (1) = 160 and T (3) = 90 then give the linear system of equations
c1 I0 (a) + c2 K0 (a) = 90
c1 I0 (3a) + c2 K0 (3a) = 20
456 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
and
x′ (t) = −0.5c1 J0′ (10e−0.05t ) − 0.5c2 Y0′ (10e−0.05t ).
c1 J0 (10) + c2 Y0 (10) = 1
c1 J0′ (10) + c2 Y0′ (10) = 1.
Using Y0′ = −Y1 and J0′ = −J1 and Table 5.2 we find c1 = −4.7860 and c2 = −3.1803.
Thus
x(t) = −4.7860J0 (10e−0.05t ) − 3.1803Y0 (10e−0.05t ).
x
(b)
10
5
t
1
46. (a) Identifying α = 2 , the general solution of x′′ + 14 tx = 0 is
1/2 1 3/2 1/2 1 3/2
x(t) = c1 x J1/3 x + c2 x J−1/3 x .
3 3
(b) x
t
50 100 150 200
−1
d2 θ
+ α2 tθ = 0, θ ′ (0) = 0, θ(L) = 0,
dt2
(b) Looking at the series forms of θ1 and θ2 we see that θ1′ (0) 6= 0, while θ2′ (0) = 0. Thus,
the boundary condition θ ′ (0) = 0 implies c1 = 0, and so
√ 2 3/2
θ(t) = c2 t J−1/3 αt .
3
so either c2 = 0, in which case θ(t) = 0, or J−1/3 ( 32 αL3/2 ) = 0. The column will just
start to bend when L is the length corresponding to the smallest positive zero of J−1/3 .
(c) Using Mathematica, the first positive root of J−1/3 (x) is x1 ≈ 1.86635. Thus
2 3/2 = 1.86635 implies
3 αL
2/3 1/3
3(1.86635) 9EI
L= = (1.86635)2
2α 4δg
1/3
9(2.6 × 107 )π(0.05)4 /4
= (1.86635)2 ≈ 76.9 in.
4(0.28)π(0.05)2
458 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
48. (a) Writing the differential equation in the form xy ′′ +(P L/M )y = 0, we identify λ = P L/M .
From Problem 37 the solution of this differential equation is
√ p √ p
y = c1 x J1 2 P Lx/M + c2 x Y1 2 P Lx/M .
√
(b) From y(L) = 0 we have y = J1 (2L P M ) = 0. The first positive zero of J1 is 3.8317 so,
p
solving 2L P1 /M = 3.8317, we find P1 = 3.6705M/L2 . Therefore,
r !
√ 3.6705x √ 3.8317 √
y1 (x) = c1 x J1 2 = c1 x J1 √ x .
L L
0.2
0.1
x
0.2 0.4 0.6 0.8 1
d2 θ dθ
(l0 + vt) 2
+ 2v + gθ = 0.
dt dt
l0 + vt d2 θ dθ g
+2 + θ = 0.
v dt2 dt v
dθ dθ dx dθ
= =
dt dx dt dx
and
d2 θ d(dθ/dt) d(dθ/dx) dx d2 θ
= = = .
dt2 dt dx dt dx2
Thus, the differential equation becomes
d2 θ dθ g d2 θ 2 dθ g
x +2 + θ=0 or + + θ = 0.
dx2 dx v dx2 x dx vx
5.3 Special Functions 459
(c) The differential equation in part (b) has the form of (18) in the text with
g
1 − 2a = 2 2c − 2 = −1 b2 c2 = a2 − p2 c2 = 0
v
r
1 1 g
a=− c= b=2 p = 1.
2 2 v
C1 u−1 −1
0 J1 (u0 ) + C2 u0 Y1 (u0 ) = θ0
C1 u−1 −1
0 J2 (u0 ) + C2 u0 Y2 (u0 ) = 0
2p
Y1 g(l0 + vt)
2p v
+J2 gl0 √ .
v l0 + vt
1 1
(e) When l0 = 1 ft, θ0 = radian, and v = 60
10 ft/s, the above function is
√ √
J1 (480 2(1 + t/60)) Y1 (480 2 (1 + t/60))
θ(t) = −1.69045 p − 2.79381 p .
1 + t/60 1 + t/60
The plots of θ(t) on [0, 10], [0, 30], and [0, 60] are
θ (t) θ (t) θ (t)
0.1 0.1 0.1
t t t
2 4 6 8 10 5 10 15 20 25 30 10 20 30 40 50 60
–0.05 –0.05 –0.05
0.05
t
50 100 150 200 250 300
–0.05
–0.1
where
1·3·5·7 35
c1 = (−1)3 =− .
2·4·6 16
Thus
35 3 99 5 429 7 1
P7 (x) = − x − 9x + x − x = (429x7 − 693x5 + 315x3 − 35x).
16 5 35 16
(b) P6 (x) satisfies 1 − x2 y ′′ −2xy ′ +42y = 0 and P7 (x) satisfies 1 − x2 y ′′ −2xy ′ +56y = 0.
That is,
d2 y dy
1 − x2 2
− 2x + n(n + 1)y = 0.
dx dx
53. The only solutions bounded on [−1, 1] are y = cPn (x), c a constant and n = 0, 1, 2, . . . . By
(iv) of the properties of the Legendre polynomials, y(0) = 0 or Pn (0) = 0 implies n must be
odd. Thus the first three positive eigenvalues correspond to n = 1, 3, and 5 or λ1 = 1 · 2,
λ2 = 3 · 4 = 12, and λ3 = 5 · 6 = 30. We can take the eigenfunctions to be y1 = P1 (x),
y2 = P3 (x), and y3 = P5 (x).
54. (a)
0 d0
P00 (x) = 1 − x2 · P0 (x) = 1
dx0
0 d0
P10 (x) = 1 − x2 · P1 (x) = x
dx0
1/2 d1 2 1/2
P11 (x) = 1 − x2 · P1 (x) = 1 − x
dx1
1/2 d1 1/2
P21 (x) = 1 − x2 · 1
P2 (x) = 1 − x2 · (3x)
dx
d2
P22 (x) = 1 − x2 · 2 P2 (x) = 1 − x2 · (3)
dx
1/2 d1 1/2 1
P31 (x) = 1 − x2 · 1
P3 (x) = 1 − x2 · 15x2 − 3
dx 2
d2
P32 (x) = 1 − x2 · 2 P3 (x) = 1 − x2 · (15x)
dx
3/2 d3 3/2
P33 (x) = 1 − x2 · 3
P3 (x) = 1 − x2 · (15)
dx
dm
(c) Because Pn (x) is a polynomial, Pn (x) = 0 for m > n. Therefore Pnm (x) = 0 for
dxm
m > n.
h i
1
(d) When n = m = 1, the associated Legendre equation is (1−x2 )y ′′ −2xy ′ + 2 − 1−x2
y=0
5.3 Special Functions 463
1/2
now force the function y = P11 (x) = 1 − x2 into the equation to get
1/2
d2 1 − x2 1/2
2 d 1 − x2 1 1/2
1−x 2
− 2x + 2− 2
1 − x2 =0
dx dx 1−x
! !
2
1 x 1 2 1/2
1−x − − 2x − + 2 − 1 − x =0
(1 − x2 )3/2 (1 − x2 )1/2 1 − x2
" #
1 2x2 2 1/2
1
− + + 2 1−x − =0
(1 − x2 )1/2 (1 − x2 )1/2 (1 − x2 )1/2
" #
1 2x2 1 − 2x2
− + + =0
(1 − x2 )1/2 (1 − x2 )1/2 (1 − x2 )1/2
0=0
1 d 1 d2 1
P1 (x) = (x2 − 1)1 = x P2 (x) = (x2 − 1)2 = (3x2 − 1)
2 dx 22 2! dx2 2
1 d3 1 1 d4 1
P3 (x) = 3 3
(x2 − 1)3 = (5x3 − 3x) P4 (x) = 4 4
(x2 − 1)4 = (35x4 − 30x2 + 3)
2 3! dx 2 2 4! dx 8
1 d5 1
P5 (x) = (x2 − 1)5 = (63x5 − 70x3 + 15x)
25 5! dx5 8
1 d6 1
P6 (x) = 6 6
(x2 − 1)6 = (231x6 − 315x4 + 105x2 − 5)
2 6! dx 16
1 d7 1
P7 (x) = (x2 − 1)7 = (429x7 − 693x5 + 315x3 − 35x)
27 7! dx7 16
P1 P2 P3 P4
56. 1 1 1 1
x x x x
–1 –0.5 0.5 1 –1 –0.5 0.5 1 –1 –0.5 0.5 1 –1 –0.5 0.5 1
–1 –1 –1 –1
P5 P6 P7
1 1 1
x x x
–1 –0.5 0.5 1 –1 –0.5 0.5 1 –1 –0.5 0.5 1
–1 –1 –1
464 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
P1 (x) : 0
P2 (x) : ± 0.57735
P3 (x) : 0, ±0.77460
P4 (x) : ± 0.33998, ±0.86115
P5 (x) : 0, ±0.53847, ±0.90618
P6 (x) : ± 0.23862, ±0.66121, ±0.93247
P7 (x) : 0, ±0.40585, ±0.74153 , ±0.94911
P10 (x) : ± 0.14887, ±0.43340, ±0.67941, ±0.86506, ±0.097391
The zeros of any Legendre polynomial are in the interval (−1, 1) and are symmetric with
respect to 0.
Chapter 5 in Review
1. False; J1 (x) and J−1 (x) are not linearly independent when ν is a positive integer. (In this
case ν = 1). The general solution of x2 y ′′ + xy ′ + (x2 − 1)y = 0 is y = c1 J1 (x) + c2 Y1 (x).
3. x = −1 is the nearest singular point to the ordinary point x = 0. Theorem 5.1.1 guarantees
X∞
the existence of two power series solutions y = cn xn of the differential equation that
n=1
converge at least for −1 < x < 1. Since − 12 ≤ x ≤ 12 is properly contained in −1 < x < 1,
both power series must converge for all points contained in − 21 ≤ x ≤ 21 .
2c2 + 2c1 + c0 = 0
6c3 + 4c2 + c1 = 0
1
12c4 + 6c3 − c1 + c2 = 0
3
2
20c5 + 8c4 − c2 + c3 = 0
3
have
1
c2 = − c0
2
2 1
c3 = − c2 = c0
3 3
1 1 1
c4 = − c3 − c2 = − c0
2 12 8
2 1 1 1
c5 = − c4 + c2 − c3 = c0 ;
5 30 20 60
whereas the assumption that c0 = 0, c1 6= 0 implies
c2 = −c1
2 1 1
c3 = − c2 − c1 = c1
3 6 2
1 1 1 5
c4 = − c3 + c1 − c2 = − c1
2 36 12 36
2 1 1 1
c5 = − c4 + c2 − c3 = − c1 .
5 30 20 360
five terms of two power series solutions are then
1 1 1 1
y1 (x) = c0 1 − x2 + x3 − x4 + x5 + · · ·
2 3 8 60
and
1 5 1 5
y2 (x) = c1 x − x2 + x3 − x4 − x + ··· .
2 36 360
5. The interval of convergence is centered at 4. Since the series converges at −2, it converges
at least on the interval [−2, 10). Since it diverges at 13, it converges at most on the interval
[−5, 13). Thus, at −7 it does not converge, at 0 and 7 it does converge, and at 10 and 11 it
might converge.
6. We have
x3 x5
sin x x−+ − ··· x3 2x5
f (x) = = 6 120 = x + + + ··· .
cos x x2 x4 3 15
1− + − ···
2 24
8. The differential equation (x − 1)(x + 3)y ′′ + y = 0 has regular singular points at x = 1 and
x = −3.
466 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
∞
X
9. Substituting y = cn xn+r into the differential equation we obtain
n=0
∞
X
2xy ′′ + y ′ + y = 2r 2 − r c0 xr−1 + [2(k + r)(k + r − 1)ck + (k + r)ck + ck−1 ]xk+r−1 = 0
k=1
which implies
2r 2 − r = r(2r − 1) = 0
and
(k + r)(2k + 2r − 1)ck + ck−1 = 0.
The indicial roots are r = 0 and r = 1/2. For r = 0 the recurrence relation is
ck−1
ck = − , k = 1, 2, 3, . . . ,
k(2k − 1)
so
1 1
c1 = −c0 , c2 = c0 , c3 = − c0 .
6 90
∞
X ∞
X ∞
X
y ′′ − xy ′ − y = n(n − 1)cn xn−2 − ncn xn − cn xn
n=2 n=1 n=0
| {z } | {z } | {z }
k=n−2 k=n k=n
∞
X ∞
X ∞
X
k k
= (k + 2)(k + 1)ck+2 x − kck x − ck xk
k=0 k=1 k=0
∞
X
= 2c2 − c0 + [(k + 2)(k + 1)ck+2 − (k + 1)ck ]xk = 0.
k=1
Chapter 5 in Review 467
Thus
and
1
c2 = c0
2
1
ck+2 = ck , k = 1, 2, 3, . . . .
k+2
∞
X
12. Substituting y = cn xn into the differential equation we obtain
n=0
∞
X
y ′′ − x2 y ′ + xy = 2c2 + (6c3 + c0 )x + [(k + 3)(k + 2)ck+3 − (k − 1)ck ]xk+1 = 0
k=1
k−1
ck+3 = ck , k = 1, 2, 3, . . . .
(k + 3)(k + 2)
c3 = c6 = c9 = · · · = 0 c4 = c7 = c10 = · · · = 0 c5 = c8 = c11 = · · · = 0
∞
X
xy ′′ −(x+2)y ′ +2y = (r 2 −3r)c0 xr−1 + [(k + r)(k + r − 3)ck − (k + r − 3)ck−1 ]xk+r−1 = 0,
k=1
which implies
r 2 − 3r = r(r − 3) = 0
and
(k + r)(k + r − 3)ck − (k + r − 3)ck−1 = 0.
Then
= 0.
Thus
1
c0 + 2c2 = 0 c1 + 6c3 = 0 12c4 = 0 20c5 − 2c3 = 0 30c6 − 5c4 + c2 = 0
12
and
1 1 1 1 1
c2 = − c0 c3 = − c1 c4 = 0 c5 = c3 c6 = c4 − c2 .
2 6 10 6 360
Choosing c0 = 1 and c1 = 0 we find
1 1
c2 = − , c3 = 0, c4 = 0, c5 = 0, c6 =
2 720
and so on. For c0 = 0 and c1 = 1 we find
1 1
c2 = 0, c3 = − , c4 = 0, c5 = − , c6 = 0
6 60
and so on. Thus, two solutions are
1 1 6 1 1
y 1 = 1 − x2 + x + ··· and y 2 = x − x3 − x5 + · · · .
2 720 6 60
470 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
∞
X
15. Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X
n−2 n
′′ ′
y + xy + 2y = n(n − 1)cn x + ncn x +2 cn xn
n=2 n=1 n=0
| {z } | {z } | {z }
k=n−2 k=n k=n
∞
X ∞
X ∞
X
= (k + 2)(k + 1)ck+2 xk + kck xk + 2 ck xk
k=0 k=1 k=0
∞
X
= 2c2 + 2c0 + [(k + 2)(k + 1)ck+2 + (k + 2)ck ]xk = 0.
k=1
Thus
2c2 + 2c0 = 0
(k + 2)(k + 1)ck+2 + (k + 2)ck = 0
and
c2 = −c0
1
ck+2 = − ck , k = 1, 2, 3, . . . .
k+1
1 1
c2 = −1 c3 = c5 = c7 = · · · = 0 c4 = c6 = −
3 15
and so on. For c0 = 0 and c1 = 1 we obtain
1 1 1
c2 = c4 = c6 = · · · = 0 c3 = − c5 = c7 = −
2 8 48
and so on. Thus, the general solution is
2 1 4 1 6 1 3 1 5 1 7
y = C0 1 − x + x − x + · · · + C1 x − x + x − x + · · ·
3 15 2 8 48
and
4 2 3 5 7
′
y = C0 −2x + x3 − x5 + · · · + C1 1 − x2 + x4 − x6 + · · · .
3 5 2 8 48
Setting y(0) = 3 and y ′ (0) = −2 we find c0 = 3 and c1 = −2. Therefore, the solution of the
initial-value problem is
1 1 1
y = 3 − 2x − 3x2 + x3 + x4 − x5 − x6 + x7 + · · · .
4 5 24
Chapter 5 in Review 471
∞
X
16. Substituting y = cn xn into the differential equation we have
n=0
∞
X ∞
X ∞
X
(x + 2)y ′′ + 3y = n(n − 1)cn xn−1 +2 n(n − 1)cn xn−2 +3 cn xn
n=2 n=2 n=0
| {z } | {z } | {z }
k=n−1 k=n−2 k=n
∞
X ∞
X ∞
X
k k
= (k + 1)kck+1 x + 2 (k + 2)(k + 1)ck+2 x + 3 ck xk
k=1 k=0 k=0
∞
X
= 4c2 + 3c0 + [(k + 1)kck+1 + 2(k + 2)(k + 1)ck+2 + 3ck ]xk = 0.
k=1
Thus
4c2 + 3c0 = 0
(k + 1)kck+1 + 2(k + 2)(k + 1)ck+2 + 3ck = 0
and
3
c2 = − c0
4
k 3
ck+2 = − ck+1 − ck , k = 1, 2, 3, . . . .
2(k + 2) 2(k + 2)(k + 1)
Setting y(0) = 0 and y ′ (0) = 1 we find c0 = 0 and c1 = 1. Therefore, the solution of the
initial-value problem is
1 1
y = x − x3 + x4 + · · · .
4 16
17. The singular point of (1 − 2 sin x)y ′′ + xy = 0 closest to x = 0 is π/6. Hence a lower bound
is π/6.
472 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
y1 = c0 [1 + · · · ] and y2 = c1 [x + · · · ],
the initial conditions at x = 1 give solutions for c0 and c1 in terms of infinite series. Letting
t = x − 1 the initial-value problem becomes
d2 y dy
2
+ (t + 1) + y = 0, y(0) = −6, y ′ (0) = 3.
dt dt
∞
X
Substituting y = cn tn into the differential equation, we have
n=0
∞ ∞ ∞ ∞
d2 y dy X
n−2
X
n
X
n−1
X
+ (t + 1) + y = n(n − 1)cn t + ncn t + ncn t + cn tn
dt2 dt n=2 n=1 n=1 n=0
| {z } | {z } | {z } | {z }
k=n−2 k=n k=n−1 k=n
∞
X ∞
X ∞
X ∞
X
= (k + 2)(k + 1)ck+2 tk + kck tk + (k + 1)ck+1 tk + ck tk
k=0 k=1 k=0 k=0
∞
X
= 2c2 + c1 + c0 + [(k + 2)(k + 1)ck+2 + (k + 1)ck+1 + (k + 1)ck ] tk = 0.
k=1
Thus
2c2 + c1 + c0 = 0
(k + 2)(k + 1)ck+2 + (k + 1)ck+1 + (k + 1)ck = 0
and
c1 + c0
c2 = −
2
ck+1 + ck
ck+2 =− , k = 1, 2, 3, . . . .
k+2
The initial conditions then imply c0 = −6 and c1 = 3. Thus the solution of the initial-value
problem is
1 2 1 3 1 4
y = −6 1 − (x − 1) + (x − 1) + (x − 1) + · · ·
2 6 12
1 2 1 3 1 4
+ 3 (x − 1) − (x − 1) − (x − 1) + (x − 1) + · · · .
2 6 6
∞
X ∞
X ∞
X
k k
= (k + 2)(k + 1)ck+2 x + (k − 1)ck−1 x + 2 ck−1 xk
k=0 k=2 k=1
∞
X
= 2c2 + (6c3 + 2c0 )x + [(k + 2)(k + 1)ck+2 + (k + 1)ck−1 ]xk = 5 − 2x + 10x3 .
k=2
and
(k + 2)(k + 1)ck+2 + (k + 1)ck−1 = 0, k = 4, 5, 6, . . . ,
Therefore
5 1 1 1 1 1 1 1 5
c2 = c3 = − c0 − c4 = − c1 c5 = − c2 = − =0
2 3 3 4 2 5 2 5 2
and
1
ck+2 = − ck−1 .
k+2
Using the recurrence relation, we find
1 1 1 1
c6 = − c3 = (c0 + 1) = 2 c0 + 2
6 3·6 3 · 2! 3 · 2!
1 1
c7 = − c4 = c1
7 4·7
c8 = c11 = c14 = · · · = 0
1 1 1
c9 = − c6 = − 3 c0 − 3
9 3 · 3! 3 · 3!
1 1
c10 = − c7 = − c1
10 4 · 7 · 10
1 1 1
c12 = − c9 = 4 c0 + 4
12 3 · 4! 3 · 4!
1 1
c13 = − c0 = c1
13 4 · 7 · 10 · 13
and so on. Thus
1 3 1 6 1 9 1 12
y = c0 1 − x + 2 x − 3 x + 4 x − ···
3 3 · 2! 3 · 3! 3 · 4!
1 4 1 7 1 10 1 13
+ c1 x − x + x − x + x − ···
4 4·7 4 · 7 · 10 4 · 7 · 10 · 13
5 2 1 3 1 6 1 9 1 12
+ x − x + 2 x − 3 x + 4 x − ··· .
2 3 3 · 2! 3 · 3! 3 · 4!
1 du
22. (a) From y = − we obtain
u dx
2
dy 1 d2 u 1 du
=− 2
+ 2 .
dx u dx u dx
(b) The differential equation u′′ + x2 u = 0 has the form (18) in the text with
1 − 2a = 0 2c − 2 = 2 b2 c2 = 1 a2 − p2 c2 = 0
1 1 1
a= c=2 b= p=
2 2 4
Then, by (19) in the text,
1/2 1 2 1 2
u=x c1 J1/4 x + c2 J−1/4 x .
2 2
(c) We have
1 du 1 d 1/2 1 1/2 dw dt 1 −1/2
y=− = − 1/2 x w(t) = − 1/2 x + x w
u dx x w(t) dx x w dt dx 2
1 dw 1 1 dw 1 dw
=− x3/2 + 1/2 w = − 2x2 +w =− 4t +w .
x1/2 w dt 2x 2xw dt 2xw dt
Now
dw d
4t + w = 4t [c1 J1/4 (t) + c2 J−1/4 (t)] + c1 J1/4 (t) + c2 J−1/4 (t)
dt dt
1 1
= 4t c1 J−3/4 (t) − J1/4 (t) + c2 − J−1/4 (t) − J3/4 (t)
4t 4t
J3/4 ( 21 x2 ) − cJ−3/4 ( 21 x2 )
y=x .
cJ1/4 ( 21 x2 ) + J−1/4 ( 21 x2 )
The new equation has the form of (12) from Section 5.3 with α = 4 and υ = 23 . Thus
the general solution is given by
(b) This equation has the parametric form of the modified Bessel equation of order υ = 3
with α = 6. Therefore the general solution is
y = c1 I3 (6x) + c2 K3 (6x)
24. (a) Using formula (5) of Section 3.2 in the text, we find that a second solution of
(1 − x2 )y ′′ − 2xy ′ = 0 is
Z R 2x dx/(1−x2 ) Z
e 2
y2 (x) = 1 · 2
dx = e− ln(1−x ) dx
1
Z
dx 1 1+x
= = ln ,
1 − x2 2 1−x
where partial fractions was used to obtain the last integral.
(b) Using formula (5) of Section 3.2 in the text, we find that a second solution of
(1 − x2 )y ′′ − 2xy ′ + 2y = 0 is
Z R 2x dx/(1−x2 ) Z − ln (1−x2 )
e e
y2 (x) = x · 2
dx = x dx
x x2
Z
dx 1 1+x 1 x 1+x
=x =x ln − = ln − 1,
x2 (1 − x2 ) 2 1−x x 2 1−x
where partial fractions was used to obtain the last integral.
(c) x y2 y2
2 2
1 1
x x
–1 1 –1 1
–1 –1
–2 –2
1 1+x x 1+x
y2 (x) = ln y2 = ln −1
2 1−x 2 1−x
25. By the binomial theorem we have
−1/2
1 + t2 − 2xt
1 2 (−1/2)(−3/2) 2 (−1/2)(−3/2)(−5/2) 2
=1− t − 2xt + (t − 2xt)2 + (t − 2xt)3 + · · ·
2 2! 3!
1 3 5
= 1 − (t2 − 2xt) + (t2 − 2xt)2 − (t2 − 2xt)3 + · · ·
2 8 16
∞
1 2 2 1 3 3
X
= 1 + xt + (3x − 1)t + (5x − 3x)t + · · · = Pn (x)tn .
2 2 n=0
Chapter 5 in Review 477
∞
1 X
(1 − 2t + t2 )−1/2 = (1 − t)−1 = = 1 + t + t2 + t3 + . . . (|t| < 1) = tn .
1−t n=0
∞
X ∞
X
n 2 −1/2
Pn (1)t = (1 − 2t + t ) = tn .
n=0 n=0
Equating the coefficients of corresponding terms in the two series, we see that Pn (1) = 1.
Similarly, letting x = −1 we have
1
(1 + 2t + t2 )−1/2 = (1 + t)−1 = = 1 − t + t2 − 3t3 + . . . (|t| < 1)
1+t
∞
X ∞
X
n n
= (−1) t = Pn (−1)tn ,
n=0 n=0
∞
X ∞
X
k
27. Force y = ck x into the equation to get [(k + 2)(k + 1)ck+2 − 2kck + 2αck ]xk = 0
k=0 k=0
therefore
(2α − 2k)
ck+2 = − ck .
(k + 2)(k + 1)
(2α) 2α
c2 = − c0 = − c0
(2)(1) 2!
(2α − 2) 2(α − 1)
c3 = − c1 = − c1
(3)(2) 3!
(2α − 4) 2(α − 2) 2α 22 α(α − 2)
c4 =− c2 = − · − c0 = c0
(4)(3) 4·3 2! 4!
(2α − 6) 2(α − 3) 2(α − 1) 22 (α − 1)(α − 3)
c5 =− c3 = − · − c1 = c1
(5)(4) 5·4 3! 5!
2
(2α − 8) 2(α − 4) 2 α(α − 2) 23 α(α − 2)(α − 4)
c6 =− c4 = − · c0 = − c0
(6)(5) 6·5 4! 6!
2
(2α − 10) 2(α − 5) 2 (α − 1)(α − 3) 23 (α − 1)(α − 3)(α − 5)
c7 =− c5 = − · c1 = − c1
(7)(6) 7·6 5! 7!
478 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
2α 2(α − 1) 22 α(α − 2)
= c0 + c1 x − c0 x2 − c1 x3 + c0 x4
2! 3! 4!
22 (α − 1)(α − 3) 23 α(α − 2)(α − 4)
+ c1 x5 − c0 x6 + . . .
5! 6!
2α 2 22 α(α − 2) 4 23 α(α − 2)(α − 4) 6
= c0 1 − x + x − x + ...
2! 4! 6!
2(α − 1) 3 22 (α − 1)(α − 3) 5
+ c1 x − x + x + ...
3! 5!
= c0 y1 (x) + c1 y2 (x)
where we take
∞
X (−1)k 2k α(α − 2)(α − 4) . . . (α − (2k − 2))
y1 (x) = 1 + x2k
(2k)!
k=1
∞
X (−1)k 2k (α − 1)(α − 3) . . . (α − (2k − 1))
y2 (x) = x + x2k+1
(2k + 1)!
k=1
H0 (x) = 20 · (1) = 1
H1 (x) = 21 · (x) = 2x
∞
X ∞
X ∞
X
(1 − x2 ) · k(k − 1)ck xk−2 − x · kck xk−1 + α2 ck xk = 0
k=2 k=1 k=0
Therefore we have
α2 1 − α2 k2 − α2
c2 = − c0 , c3 = c1 , and ck+2 = ck for k = 2, 3, 4, . . .
2! 3! (k + 2)(k + 1)
Therefore we get
∞
X
y= ck xk = c0 + c1 x + c2 x2 + c3 x3 + c4 x4 + c5 x5 + c6 x6 + . . .
k=0
α2 1 − α2 (4 − α2 )α2 (9 − α2 )(1 − α2 )
= c0 + c1 x − c0 x2 + c1 x3 − c0 x4 + c1 x5
2! 3! 4! 5!
(16 − α2 )(4 − α2 )α2
− c0 x6 + . . .
6!
α2 2 α2 (4 − α2 ) 4 (16 − α2 )(4 − α2 )α2 6
= c0 1 − x − x − x + ...
2! 4! 6!
1 − α2 3 (9 − α2 )(1 − α2 ) 5
+ c1 x + x + x + ...
3! 5!
= c0 y1 (x) + c1 y2 (x)
480 CHAPTER 5 SERIES SOLUTIONS OF LINEAR DIFFERENTIAL EQUATIONS
where we take
α2 2 (4 − α2 )α2 4 (16 − α2 )(4 − α2 )α2 6
y1 (x) = 1 − x − x − x + ...
2! 4! 6!
1 − α2 3 (9 − α2 )(1 − α2 ) 5
y2 (x) = x + x + x + ...
3! 5!