The LNM Institute of Information Technology

Jaipur, Rajasthan

Second Mid Semester Examination

.

Mathematics-II, 11th March 2014

Time: 60 Minutes, Maximum Marks: 40

Note: Attempt all the questions. Make an index showing the question number attempted on the
page number in top of the answer sheet in the following format, otherwise you may be penalized
by 2 marks. Read the questions carefully. No query will be entertained after 15 minutes of
Examination.
Question No.
Page No.

1. (a) Find a solution for the dierential equation (1 + x2 )y 00 4xy 0 + 6y = 0 such that y =
a0 y1 (x) + a1 y2 (x), where y1 and y2 are power series.
[5]

Solution:

Substituting y =

an xn into (1 + x2 )y 00 4xy 0 + 6y = 0, we get an+2 =

n=0
(n2)(n3)
(n+1)(n+2) an ;

n 0: Iterating we get a2 = 3a0 ; a3 = 13 a1 ; an = 0; n 4. Thus,

y = a0 (1 3x2 ) + a1 (x 13 x3 ) = a0 y1 (x) + a1 y2 (x).
1 dn
(b) Prove the Rodriguess formula: Pn (x) = n
[(x2 1)n ].
[5]
2 n! dxn
Solution: Let V (x) = (x2 1)n . Then V 0 (x) = 2nx(x2 1)n1 or (x2 1)V 0 (x) =
2nxV (x). Now dierentiating (n + 1) times (by the use of the Leibniz rule for dierentiation), we get (x2 1)V (n+2) (x) + 2(n + 1)xV (n+1) (x) + 2n(n+1)
V (n) (x) 2nxV (n+1) (x)
2
2n(n + 1)V (n) (x) = 0. By denoting, U (x) = V (n) (x), we have or (1 x2 )U 00 2xU 0 +
n(n + 1)U = 0. This tells us that U (x) is a solution of the Legendre Equation. Also, let
n
dn
dn
2
n
n
us note that dx
n (x 1) = dxn (x 1)(x + 1) = n!(x + 1) + terms containing a factor

n
d
2
n
of (x 1). Therefore, dx
= 2n n!.
n (x 1) 
1 dn
[(x2 1)n ]
So, Pn (x) = n
2 n! dxn

x=1

2. (a) Reduce x2 y 00 + xy 0 + (x2 1/4)y = 0 to normal form and hence nd its general solution.
[5]
R

Solution: Here p = x1 ; q = 1 4x12 . So v = e 2 pdx = 1x . Using the transformation

y = uv, we get u00 + u = 0 and its general solution is u = A sin x + B cos x. For the
original equation, the general solution is y = 1x (A sin x + B cos x).
1

(b) Prove that between each pair of consecutive positive zeros of J (x), there is exactly one
zero of J+1 (x) and vice versa.
[5]
Solution: Let a and b be two consecutive positive zeros of J+1 . Let f (x) = x+1 J+1 (x).
Then f (a) = f (b) = 0. Thus there exists c between a and b such that f 0 (c) = 0. So
f 0 (c) = (c+1 J+1 (c))0 = c+1 J (c) = 0. Thus J (c) = 0.

Similarly, let a and b be two consecutive positive zeros of J . Let f (x) = x J (x).
Then f (a) = f (b) = 0. Thus there exists c between a and b such that f 0 (c) = 0. So
f 0 (c) = (c J (c))0 = c J+1 (c) = 0. Thus J+1 (c) = 0.
To prove uniqueness, let there exist two zeros c and d of J (x) between consecutive zeros
a and b of J+1 (x) This implies that there exist a zero of J+1 (x) between c and d.
Consequently, there is a zero of J+1 (x) between a and b, Which contradicts the fact
that a and b are consecutive zeros of J+1 (x).
3. (a) Find all the nontrivial
solutions of the boundary value problem y 00 + y = 0 for x (0, 2 )
(
)
with y(0) = 0, y 0 2 = 0. For any piecewise smooth function f (x) dened on [0, 2 ], nd
cm in terms of f (x) such that
f (x) =

cm ym (x),

m=1

where ym (x) is a solution of given boundary value problem.

[6]
Sol. Case I: = 0 gives y(x) = Ax + B. Boundary Condition implies y 0.
Case II: Let < 0 and = n2 . The solution is y(x) = Aenx + Benx . Boundary
condition implies y 0.
Case III: Let > 0 and = n2 . The solution is y(x) = A cos nx + B sin nx.
y(0) = 0 A = 0. y 0 (/2) Bn cos n
2 = 0. B = 0 implies y 0, so assume B 6= 0.
n
2m1
n
Hence, cos 2 = 0 and 2 = 2 , m = 0, 1, 2, .
The eigen values are m = n2 = (2m 1)2 , m = 1, 2, .
By taking B = 1, the corresponding eigen functions are
ym (x) = sin(2m 1)x, m = 1, 2, 3,
For anypiecewise smooth function f (x) dened on [0, 2 ], we can represent f (x) as
f (x) =
m=1 cm ym (x).
Multiply ym both sides and integrate from 0 to /2 w.r.t. x. Since eigenfunctions are
orthogonal we get
/2
cm =

f (x) sin(2m 1)xdx

/2
0

sin2 (2m 1)xdx

/2

f (x) sin(2m 1)xdx.

(b) Let {ym }

m=1 be the eigen functions of the following Sturm-Liouville problem:
[
]
d
dy
p(x)
+ [q(x) + r(x)]y = 0, a < x < b
dx
dx
with boundary condition:
c1 y(a) + c2 y 0 (a) = 0

and

d1 y(b) + d2 y 0 (b) = 0,

where c1 or c2 6= 0 and d1 or d2 6= 0, p(x) > 0 and q(x) > 0. Show that eigen functions
{ym }
[4]
m=1 are orthogonal with respect to the weight function q(x).
Sol. Assume ym and yn are two distinct eigen functions corresponding to two distinct eigen
values m and n . This implies
0 0
[pym
] + [m q + r]ym = 0,

(1)

[pyn0 ]0 + [n q + r]yn = 0.

(2)

and

Multiplying yn into (1), ym into (2) and subtracting we get

0 0
yn [pym
] ym [pyn0 ]0 + (m n )qym yn = 0.

(3)

Integrating (3) from a to b w.r.t x, we obtain

b
(m n )
qym yn dx = p(b)W (b) p(a)W (a).
a

Since ym and yn satises the boundary conditions, we get

0
c1 ym (a) + c2 ym
(a) = 0

and

c1 yn (a) + c2 yn0 (a) = 0.

c1 and c2 cant be zero simultaneously, so it has a nontrivial solution c1 and c2 which implies
the determinant of coecient matrix must be zero. Therefore, W (a)=0. A similar argument
shows that W (b) = 0.
b
Since m 6= n , a qym yn dx = 0 and hence eigen functions are orthogonal.
4. (a) Using Laplace transformation determine the hammerblow response of the following
damped mass-spring system under the unit impulse at time t = 1:
[5]
y 00 + 2y 0 3y = (t 1),
y(0) = 0, y 0 (0) = 2.
Sol. Let L{y(t)} = Y (s).
L{y 0 (t)} = sY y(0) = sY and L{y 00 (t)} = s2 Y sy(0) y 0 (0) = s2 Y 2
Taking Laplace transformation of the given DE, we get
s2 Y 2 + 2sY 3Y = es .
Solving for Y , we get
Y =

es + 2
.
(s 1)(s + 3)

Using partial fraction we obtain

Y =

es
1
1
es

.
4(s 1) 4(s + 3) 2(s 1) 2(s + 3)

Applying second shifting theorem we nd the inverse Laplace transformation as

1
1
1
1
y(t) = u(t 1)e(t1) u(t 1)e3(t1) + u(t)et u(t)e3t
4
4
2
2
which is equal to
{
f (t) =

1 t
2 (e
1 t
2 (e

e3t )
0<t1
1 t1
3t
3(t1)
e ) + 4 (e
e
) t > 1.

(b) Represent f (t) = 12 te3t , 0 < t < 2 in terms of unit step function and then nd the
Laplace transorm of the resulting function.
[5]
1
3t
Sol. Unit step representation of f (t) is 2 [u(t) u(t /2)]te .
Using rst and second shifting theorem, we get
L{u(t)te3t } =

1
.
(s 3)2

L{u(t /2)te3t } = e 2 s L{(t + /2)e3(t+/2) }

(4)
(5)

Since
L{(t + /2)e3(t+/2) } = L{e

3
2

te3t +

3 3t
e 2 e },
2

applying rst shifting theorem, we nd

L{e

3
2

3
e2
e2
te + e 2 e3t } =
+
2
2
(s 3)
2 (s 3)
3t

and hence

(
s
2

L{u(t /2)te } = e
3t

e2
e2
+
2
(s 3)
2 (s 3)

)
.

Therefore, from (4)-(5) the laplace transformation of given function is:

(
[
)]
3
3
e2
e2
1
1
s
e 2
+
L{f (t)} =
.
2 (s 3)2
(s 3)2
2 (s 3)
End of paper