SOLUTION SET VI FOR 18.

075FALL 2004

4. Series Solutions of Differential Equations:Special Functions 4.2. Illustrative examples. . 5. Obtain the general solution of each of the following dierential equations in terms of Maclaurin series: d2 y = xy , (a) dx 2 (b)
d2 y dx2 dy + x dx y = 0.

Solution. (a) Try the Maclaurin series y = xy =

an xn+1 =

n=0

n=0

n=0 an x

to get a1 = 0,

an1 xn ,

The dierential equation yields [(n + 2)(n + 1)an+2 an1 ]xn = 0,

n=0

d2 y n2 n ( n 1) a x = (n + 2)(n + 1)an+2 xn . = n dx2

n=2 n=0

which is satised by all x in some neighborhood of x 0 = 0. Hence, the recurrence formula (relation) for the coecients an reads (n + 2)(n + 1)an+2 = an1 ; Find the coecients explicitly for various n: n=0: n=1: n=2: n=3: n=4: n=5: a2 = 0 3 2a3 = a0 4 3a4 = a1 a1 = 0, n = 0, 1, 2, 3, . . . .

5 4a5 = a2

6 5a6 = a3

n = 6 : 8 7a8 = a5 , . . . . Notice that a0 and a1 are independent and arbitrary, while all coecients a 2 , a5 , a8 , . . . a3n+2 . . . = 0.
Date : October 22, 2002.
1

7 6a7 = a4

SOLUTION SET VI FOR 18.075FALL 2004

The corresponding power series for y (x) reads as

x4 x7 x3n+1 + + ... + + ... y (x) = a1 x + 4 3 (3 4)(6 7) (3 4)(6 7) . . . [3n (3n + 1)]
x3 x6 x3n + + ... + + ... . +a0 1 + 2 3 (2 3)(5 6) (2 3)(5 6) . . . [(3n + 2)(3n + 3)] n (b) Once again, we try the Maclaurin series y (x) n=0 an x to get x which in turn lead to the equation [(n + 2)(n + 1)an+2 + (n 1)an ]xn = 0,
n=0

dy nan xn , = dx
n=0

d2 y = (n + 2)(n + 1)an+2 xn , 2 dx
n=0

satised by all x in some neighborhood of x 0 = 0. It follows that Write the ensuing coecients explicitly: n=1: n=2: n=3: n=4: (n + 2)(n + 1)an+2 = (n 1)an , n=0:

n = 0, 1, 2, 3, . . . .

2a2 = a0 , 4 3a4 = a2 ,

3 2a3 = 0 a1 = 0, 5 4a5 = 2a3 = 0, 6 5a6 = 3a4 ,

n = 5 : 7 6a7 = 4a5 = 0. It follows that a0 and a1 are independent and arbitrary. Further, all coecients with odd index are zero, with the exception of a 1 (since the right-hand side of the equation for n = 1 vanishes). The nal Maclaurin series for y (x) reads as
x2 x4 1 3x6 1 3 5x8 y (x) = a0 1 + + ... + 2! 4! 6! 8! 2n+2 1 3 . . . (2 n 1) x + . . . + a1 x. +(1)n (2n)! Notice that the independent solution involving a 1 is u(x) = x. 6. For each of the following equations, obtain the most general solution which is representable by a Maclaurin series: d2 y d2 y (b) dx (a) dx 2 + y = 0, 2 (x 3)y = 0, d2 y dy dy 1 2 d2 y 2 (d) x dx2 dx + y = 0, (c) 1 2 x dx2 + x dx y = 0,

SOLUTION SET VI FOR 18.075FALL 2004 d y dy 2 (e) (x2 + x) dx 2 (x 2) dx (x + 2)y = 0. Obtain three nonvanishing terms in each innite series involved. n Solution. (a) With y (x) = n=0 An x , the recurrence formula for the coecients A n is
2

(n + 2)(n + 1)An+2 + An = 0, n=0:

n = 0, 1, 2, 3, . . . .

Specically, A0 , 21 A1 n = 1 : 3 2A3 + A1 = 0 A3 = , 23 A0 A2 n = 2 : 4 3A4 + A2 = 0 A4 = = , 34 4! A3 A1 n = 3 : 5 4A5 + A3 = 0 A5 = = , ... 54 5! 2 1A2 + A0 = 0 A2 =

It follows that

x 2 x4 x 3 x5 + . . . + A1 x + ... . y (x) = A0 1 2! 4! 3! 5! n n (b) Again, start with y (x) = n=0 An x and xy (x) = n=0 An1 x , where A1 = 0, to arrive at the recurrence formula Specically, (n + 2)(n + 1)An+2 An1 + 3An = 0; 2 1A2 + 3A0 = 0 A2 = A1 = 0, n = 0, 1, 2, . . . .

n=0: n=1: n=2: It follows that

3 A0 , 12 A1 A0 3 2A3 A0 + 3A1 = 0 A3 = , 23 2 A2 A1 3A0 A1 4 3A4 A1 + 3A2 = 0 A4 = = + , ... . 34 4 34 8

3 A0 A1 A1 3A0 3 2 x + + x4 + . . . y (x) = A0 + A1 x A0 x + 2 6 2 12 8

x3 x4 3 2 1 3 + ... . = A0 1 x + x . . . + A1 x 2 6 2 12 n (c) With y (x) = n=0 An x , we get and we nd the recurrence formula dy nAn xn , = x dx n=0

2d

2y

dx2

n=0

n(n 1)An xn ,

1 (n + 2)(n + 1)An+2 (n 1)(n 2)An = 0. 2

SOLUTION SET VI FOR 18.075FALL 2004

Try dierent values of n: n=0: n=1: n=2: n=3: n=4: n=5: 2 1A2 A0 = 0 A2 = A0 , 2

It follows that all coecients An with n 3 vanish! Hence,

x2 y (x) = A0 1 + + A1 x. 2 (d) Clearly, dy = (n + 1)An+1 xn , dx n=0 x2 The recurrence formula is Specically,
n=0

3 2A3 0 = 0 A3 = 0, 4 3A4 = 0, 5 4A5 = A3 = 0, 6 5A6 3A4 = 0 A6 = 0, 7 6A7 2 3A5 = 0 A7 = 0

etc.

d2 y = n(n 1)An xn . dx2 n = 0, 1, 2, . . . ..

[n(n 1) + 1]An = (n + 1)An+1 , n=0: n=1: n=2:

A 0 = A1 , A0 , 2

A1 = 2A2 A2 =

3A2 = 3A3 A3 =

Hence, y (x) = A0 1 + x + (e) Clearly, (x + 2)y =

2 n=0

x2 + ... . 2

A0 etc. 2 A1 = 0,

An1 xn + 2

n=0

By putting all these terms together, the recurrence formula reads (n 2)(n + 1)An + (n + 1)(n + 2)An+1 nAn1 = 0;

d2 y n(n + 1)An+1 xn . n(n 1)An xn + (x + x) 2 = dx

2 n=0 n=0

dy n (n + 1)An+1 xn , (x 2) = (n 1)An1 x 2 dx n=0 n=0

An xn ,

A1 = 0,

n = 0, 1, 2, . . . .

SOLUTION SET VI FOR 18.075FALL 2004

Specically, n=0: n=1: Finally, y (x) = A0

x2 1+x+ + ... . 2 2A0 + 1 2A1 = 0 A0 = A1 , A0 2 etc.

2A1 + 2 3A2 A0 = 0 A2 =

4.3. Singular points of linear, second-order dierential equations. . 8. Locate and classify the singular points of the following dierential equations: (a) (x 1)y + xy = 0 (x 0), (b) y + y log x + xy = 0 (x 0), (c) xy + y sin x = 0, 2 (d) y |1 x |y = 0, (e) y + y cos x = 0 (x 0). Solution. (a) The singular points are x = 1 and x = 0. x = 1 is a regular singular point x x 2 2 since (x 1) (x1) = (x 1) x has a Taylor expansion near x = 1. Since (x (x1) ) |x=0

x does not exist, x2 (x 1) does not have a Taylor expansion near x = 0. So x = 0 is a irregular singular point. (b) The singular point is x = 0, which is irregular since x log x is not dierentiable at x = 0. x n1 x2n2 .) (c) There are no singular points. (Note that sin n=1 (1) x = (2n1)! (d) The singular points are x = 1 and x = 1. Since neither ((x 1) 2 |1 x2 |) |x=1 nor ((x + 1)2 |1 x2 |) |x=1 is well dened, both singular points are irregular. xn (e) There are no singular points. (Note that cos x = n=0 (1)n (2 n)! .)

4.4. The Method of Frobenius. . 11. Use the method of Frobenius to obtain the general solution of each of the following dierential equations, valid near x = 0: (a) 2xy + (1 2x)y y = 0, 1 )y = 0, (b) x2 y + xy + (x2 4 (c) xy + 2y + xy = 0, (d) x(1 x)y 2y + 2y = 0. Solution. (a)Rewrite the equation as y + 1 1 1 x ( x)y + 2 ( )y = 0. x 2 x 2

SOLUTION SET VI FOR 18.075FALL 2004

Then we can see that P0 = 1/2, P1 = 1,Q1 = 1/2,and all other Pn s, Qn s and Rn s are zeros. So f (s) = s2 1 1. f (s) = 0 has two roots: 2 s, g1 (s) = s + 1/2, and gn (s) = 0 if n = An1 1 0 s = 2 and s = 0. Take s = 0, then An = n , for all n 1. Hence, by induction, A n = A n! for all n 0. Therefore xn y = A0 = A 0 ex n ! n=1
n1 Now, take s = 1/2, then An = 2 A 2n+1 , for all n 1. Therefore 1

y = x2

n=0

An xn

= x 2 A0 Here (2n + 1)!! = 3 5 7... (2n + 1.) The general solution is then of the form: y (x) = C1 e + C2 x (b) Rewrite the equation as y +
x

2n xn . (2 n + 1)!! n=0

1 2

n=0

2n xn (2n + 1)!!

induction that A2n =

1 1 1 y + 2 (x2 )y = 0. x x 4 1 Then we can see that P0 = 1, Q0 = 4 , Q2 = 1, and all other Pn s, Qn s and Rn s are 1 1 , g2 (s) = 1, and gn (s) = 0 if n = zeros. So f (s) = s2 4 2. f (s) = 0 has two roots: s = 2 1 . and s = 2 1 For s = 2 we have An = n(n1 1) An2 for all n 2. From this, it easy to check by
(1)n (2n)! A0 1 2

and A2n+1 = An xn

(1)n (2n+1)! A1

for all n 0. So, in this case,

y = x

n=0

= A0 x

1 2

n=0
1

= A0 x 2 cos x + A1 x 2 sin x. The general solution is then of the form

1 1

(1)n

(2n)!

x ) + A1 x
1

2n

1 2

(1)n 2n+1 x ) ( (2n + 1)! n=0

y = c0 x 2 cos x + c1 x 2 sin x. (c) Rewrite the equation as 2 x2 y + 2 y = 0. x x The we can see that P0 = 2, Q2 = 1, and all other Pn s, Qn s and Rn s are zeros. So 2. f (s) = 0 has two roots: s = 1 and s = 0. f (s) = s2 + s, g2 (s) = 1, and gn (s) = 0 if n = y +

SOLUTION SET VI FOR 18.075FALL 2004

For s = 1, we have An = n(n1 1) An2 . So A2n = all n 0. Then 1 y = x An xn

n=0

(1)n (2n)! A0

and A2n+1 =

(1)n (2n+1)! A1

for

= x

= x1 (A1 sin x + A0 cos x). The general solution is then of the form

(1)n 2n (1)n 2n+1 x + A0 x ) (A1 (2n + 1)! (2n)! n=0 n=0

y = x1 (c1 sin x + c0 cos x). (d) Rewrite the equation as 2 2x y + 2 y = 0. x x Then we can see R1 = 1, P0 = 2, Q1 = 2, and all other Pn s, Qn s and Rn s are zeros. So f (s) = s2 3s, g1 (s) = s2 + 3s, and gn (s) = 0 for all n > 1. f (s) has two roots: s = 3 and s = 0. 1 (n) For s = 0, An = g f (n) An1 = An1 for all n 1, n = 3. Thus, A2 = A1 = A0 , and A3 = A4 = A5 = . So, in this case, (1 x)y y = x
0 n=0

An xn
2 3

= A0 (1 + x + x ) + A3 x x3 1 x3 = A0 + A3 . 1x 1x The general solution is then of the form y = c0

xn

n=0

1 x3 + c1 . 1x 1x

12. Use the method of Frobenius to obtain the general solution of each of the following dierential equations, valid near x = 0: (a) x2 y 2xy + (2 x2 )y = 0, (b) (x 1)y xy + y = 0, (c) xy y + 4x3 y = 0, (d) (1 cos x)y sin xy + y = 0. Solution. (a)Rewrite the equation as y 1 2 y + 2 (2 x2 )y = 0. x x

SOLUTION SET VI FOR 18.075FALL 2004

Then we can see that P0 = 2, Q0 = 2,Q2 = 1 and all other Pn s, Qn s and Rn s are 2. f (s) = 0 has two roots: zeros. So f (s) = s2 3s + 2, g2 (s) = 1, and gn (s) = 0 if n = s = 1 and s = 2. For s = 1, we have An = An2 n(n 1)

A1 A0 for n 2. From this, its easy to check by induction that A 2n = (2 n)! and A2n+1 = (2n+1)! for all n 0. So
1 1 y = x A0 x2n + A1 x2n+1 = x(A0 cosh (x) + A1 sinh (x)). (2n)! (2n + 1)! n=0 n=0

The general solution is then of the form (b) Rewrite the equation as

y = c0 x cosh (x) + c1 x sinh (x). x2 y = 0. x2 Then we can see that R1 = 1, P2 = 1, Q2 = 1, and all other Pn s, Qn s and Rn s are zeros. So f (s) = s2 s, g1 (s) = (s 1)(s 2), g2 (s) = s 3, and gn (s) = 0 if n 3. f (s) = 0 has two roots: s = 0 and s = 1. For s = 0, we have (1 x)y + xy An = n2 n3 g1 (n)An1 + g2 (n)An2 = An1 An2 n(n 1) f (n) n

0 for n 2. From this, its easy to check by induction that A n = A n! if n 2. So

xn y = A0 1 + + A1 x = A0 (ex x) + A1 x = A0 ex + (A1 A0 )x. n ! n=2

Hence the general solution is of the form

y = c0 ex + c1 x. (c) Rewrite the equation as 1 4x4 y + 2 y = 0. x x Then we can see that Q4 = 4, P0 = 1, and all other Pn s, Qn s and Rn s are zeros. So f (s) = s2 2s, g4 (s) = 4, and gn (s) = 0 if n = 4. f (s) = 0 has two roots: s = 0 and s = 2. For s = 0, we have A1 = A3 = 0, and An = n(n4 2) An4 for all n 4. From these, its y
n n

1) (1) easy to check by induction that A2n+1 = 0, A4n = ((2 n)! A0 , and A4n+2 = (2n+1)! A2 for all n 0. So (1)n 4n (1)n 4n+2 y = A0 x + A2 x = A0 cos (x2 ) + A2 sin (x2 ). (2n)! (2n + 1)! n=0 n=0

SOLUTION SET VI FOR 18.075FALL 2004

The general solution is then of the form y = c0 cos (x2 ) + c1 sin (x2 ). (d) Rewrite the equation as

(1)n 1 (1)n+1 2n 2 2n 2 x y + 2 x y + 2 y = 0. (2 n + 2)! x (2 n + 1)! x n=0 n=0
n+1 n

(1) 1) Then we can see that Q0 = 2, P2n = 2 ((2 n+1)! , R2n = 2 (2n+2)! for all n 0, and all other Pn s, Qn s and Rn s are zeros. So f (s) = (s 1)(s 2), and g 2n1 (s) = 0, g2n (s) = (1)n 2 (2 n+2)! (s 2n)(s 4n 3) for all n 1. f (s) = 0 has two roots: s = 1 and s = 2. For s = 1,using the equation n f (s + n)An = gk (s + n)Ank , k =1

its easy to check by induction that A 2n = So y = x An xn

n=0

(1)n (2n+1)! A0 ,

(1) and A2n+1 = 2 (2 n+2)! A1 for all n 0.

n

(1)n 2n (1)n 2n+1 = A0 x x + A1 x 2 x (2n + 1)! (2n + 2)! n=0 n=0 (1)n 2n+2 (1)n 2n+1 2 x x + A1 = A0 (2n + 1)! (2n + 2)! n=0 n=0

The general solution is then of the form

= A0 sin x + 2A1 (1 cos x).

y = c0 sin x + c1 (1 cos x).