Sunteți pe pagina 1din 3

P∞

1. The radius of convergence of the power series n


n=0 an x , where an = 3−n for n even and
an = 3−n+3 for n odd, is
(A) 3 (B) 1/3 (C) 27 (D) 1/27
1
Solution. Applying the root test, we get limn→∞ |an | n = 3−1 . Thus the radius of
convergence is 3.

2. Let a, b, c be the smallest positive zeroes of the Bessel functions J0 (x), J1 (x), J2 (x),
respectively. Then
(A) b < a < c (B) a < b < c (C) a < c < b (D) b < c < a
Solution. From the graphs of the Bessel functions, we see that the value of x at the first
positive intersection with the x-axis increases as the order of the Bessel function increases.

3. Let sn be the number of roots of Pn (x) in [−1, 1] and rn be the number of nonzero roots
of Pn (x) in [−1, 1]. Then it is not true that
(A) r2 + r3 = r5 (B) r3 + r5 = r8 (C) s2 + s3 = s5 (D) s3 + s5 = s8
Solution. Note that r2 = 2, r3 = 2, r5 = 4, r8 = 8 and sn = n. On substituting in the
options, we obtain: (A) 2 + 2 = 4, (B) 2 + 4 6= 8, (C) 2 + 3 = 5, (D) 3 + 5 = 8.

4. While solving the equation 2y 00 − 5xy 0 − 10y = 0 by the power series method around the
point x = 0, the recursion obtained is
2an 2an 5an 5an
(A) an+1 = 5(n+1) (B) an+2 = 5(n+1) (C) an+1 = 2(n+1) (D) an+2 = 2(n+1)

Solution. We get the relation ∞ n


P
n=0 (2(n + 1)(n + 2)an+2 − 5nan − 10an )x = 0 which
5(n+2)an 5an
gives an+2 = 2(n+1)(n+2) = 2(n+1) .

5. An inner product on R2 can be defined by setting h(a1 , a2 ), (b1 , b2 )i equal to


(A) 3a1 b1 − a2 b2 + a1 b2 + a2 b1 (B) a1 b1 + 3a2 b2 − 2a1 b2 − 2a2 b1
(C) 3a1 b1 + a2 b2 + a1 b2 + 2a2 b1 (D) 3a1 b1 + a2 b2 − a1 b2 − a2 b1
Solution. The third map is not symmetric. For the first map, ha, ai < 0 for a = (0, 1)
and for the second ha, ai = 0 for a = (1, 1). The last map is indeed an inner product since
it is symmetric and ha, ai = (a1 − a2 )2 + 2a21 > 0 for all a 6= (0, 0).

6. Let y(x) = ∞ n 2 00 0
P
n=0 an x be the solution of the equation (1 − x )y − 2xy + 20y = 0 on the
0
interval (−1, 1) with initial conditions y(0) = 0 and y (0) = 1. Then for all k > 2,
(A) |a2k +a2k+1 | = 1 (B) |a2k +a2k+1 | > 1 (C) |a2k −a2k+1 | > 0 (D) |a2k −a2k+1 | = 0
Solution. This is the Legendre equation for n = 4. Thus, y(x) = c1 P4 (x) + c2 Q4 (x).
From the initial conditions c1 = 0 since P4 (0) 6= 0 and Q4 (0) = 0. Hence y(x) is a scalar
multiple of Q4 (x). So, its power series expansion has only odd degree terms. Also it is
easily checked that the coefficients of x2k+1 are all not 1 or greater than 1 for k > 2.

7. The domain of analyticity of a real-valued function on R cannot be


(A) (0, 1) ∪ (2, 3) (B) R \ {0} (C) (−1, 1) \ {0} (D) R \ (0, 1)
Solution. The domain of analyticity is always an open set. Only the marked option is
not an open set.
8. Let m be the number of local maxima and n be the number of zeroes of J1/2 (x) in the
interval (0, 111). Then m + n equals
(A) 53 (B) 54 (C) 70 (D) 72
q
2
Solution. Observe that 35π < 111 < 36π. Starting at x = π, J1/2 (x) = πx sin x has
zeroes displaced by π, and hence, J1/2 (kπ) = 0 for k = 1, . . . , 35. Thus, n = 35. Further,
there is a local maxima for some x ∈ (2kπ, (2k + 1)π) for k = 0, . . . , 17. Thus, m = 18.

Z 1 n p
X 2
9. Suppose j(2j + 1)Pj (x) dx = 812. Then the value of n is
−1 j=1

(A) 14 (B) 27 (C) 28 (D) 29


p
Solution. Let aj = j(2j + 1). By orthogonality of Pj (x), the only terms that survive
2 P (x) 2 . Thus the lhs equals
 Pn 2
are the
Pn ones that come from aj j j=1 2aj /(2j + 1) implying
that j=1 2j = 812. This gives n = 28.

10. Let α be the smallest value of x for which the function f (x) = x8 J8 (x) defined for x > 0
has a local maxima. Then J8 (α) equals
(A) J7 (α) (B) −J7 (α) (D) −J6 (α)
(C) J6 (α)
h i
d
Solution. From the first identity, we have dx xp Jp (x) = xp Jp−1 (x). For any local
extrema α of f (x), we have f 0 (α) = α8 J7 (α) = 0. Invoking the identity Jp−1 (x) +
Jp+1 (x) = 2p
x Jp (x) yields J6 (α) + J8 (α) = 0, or J8 (α) = −J6 (α).

00 + (4x2 − 8x)y 0 + (5 − 4x)y = 0 on x > 0 has a solution of the form


11. The equation 4x2 yP
y(x) = Kx (1 + ∞
5/2 n 1/2 (1 +
P∞ n
n=1 an x ) log x + x n=1 An x ), with K equal to
1 1 1
(A) 0 (B) 4 (C) 8 (D) − 16
Solution. Rewrite the ODE as x2 y 00 + (x − 2)xy 0 + ( 54 − x)y = 0 and observe that the
indicial equation is r2 − 3r + 54 which has roots 52 and 12 that differ by N = 2. We know
that one of the solutions is y1 (x) = x5/2 (1 + n≥1 an xn ) where an satisfy the recurrence
P

1 5
(n + r − )(n + r − )an = −(n + r − 2)an−1 , n ≥ 1.
2 2
This implies
r(r − 1)
aN = a2 = 3 .
(r + 2 )(r − 12 )(r + 12 )(r − 32 )
Finally, the coefficient of the log term in the second independent solution y2 (x) is given
by
1 1
K = lim a2 (r − ) = .
1
r→ 2 2 8

12. The equation x2 (x − 1)y 00 + (1 − x2 )y 0 + xy = 0 has


(A) regular singular points at both 0 and 1
(B) regular singular point at 0 and irregular singular point at 1
(C) irregular singular point at 0 and regular singular point at 1
(D) irregular singular points at both 0 and 1
Solution. On writing in standard form, we get y 00 − x+1 x2
y 0 + x(x−1)
1
y = 0. At x = 0, the
0 x+1
coefficients of both y and y blow up at x = 0. But x2 is not analytic when multiplied
by x since it has a second order pole at x = 0. Therefore, we have an irregular singular
point at x = 0. For x = 1, the coefficient of y 0 is analytic while that of y blows up. But
when multiplied by (x − 1)2 it becomes analytic, hence it is a regular singular point.

13. While solving the equation x2 (1 + x2 )y 00 + x(4 + 2x)y 0 + y = 0 by the Frobenius method
around the point x = 0, the indicial equation obtained is
(A) r2 +3r+1 = 0 (B) r2 +7r+4 = 0 (C) r2 +15r+1 = 0 (D) r2 +31r+4 = 0
4+2x
Solution. Indicial equation is r(r − 1) + b0 r + c0 where B(x) = 1+x2
= 4 + 2x − 4x2 − ...
1 2 4
and C(x) = 1+x 2 = 1 − x + x − ....

14. Let ∞ x
P
n=0 cn Pn (x) be the Fourier-Legendre series of the function e on the interval [−1, 1].
Then (c0 , c1 ) equals
(A) ( 21 (e − 1e ), 3e) (B) ( 12 (e − 1e ), 3e ) (C) ( 12 (e − 1e ), 3(e + 1e )) (D) (1, 1)
1
Solution. We know cn = n + 21 −1 ex Pn (x)dx, for n ≥ 0. Calculation for c0 is straight-
 R

forward as P0 (x) = 1; for c1 , use P1 (x) = x and integration by parts to obtain xex − ex
as an antiderivative of xex .

15. The function x3 J4 (x) for x > 0 equals


(A) 48 − 8x2 J1 (x) − 24x − x3 J0 (x)
 
(B) 96J1 (x) − 16xJ0 (x)
(C) 96 − 8x2 J1 (x) − 16x − x3 J0 (x) (D) 24J1 (x) − 8xJ0 (x)
2p
Solution. Use the recursion x Jp (x) = Jp−1 (x) + Jp+1 (x) for p = 1, 2, and 3 successively.

16. Let y(x) be the solution of the equation x2 y 00 + (5x − 2x2 )y 0 + (4x + 3)y = 0 for x > 0
with initial conditions y(1) = 1 and y 0 (1) = 2. The least positive integer n such that
limx→∞ y(x)
xn = 0 is
(A) 2 (B) 3 (C) 4 (D) 5
Solution. The indicial equation is I(r) = r2 + 4r + 3 which has roots −3 and −1. The
question indicates that the solution involves finitely many powers of x, let us start with
the general solution corresponding to the bigger root, y1 (x) = x1 (1 + a1 x + a2 x2 + . . .).
2(n−4)
We get I(n − 1)an + an−1 (4 − 2(n − 2)) = 0 which shows an = I(n−1) an−1 . This gives,
2 2 1
y1 (x) = − 15 x + x − 2 + x . We see that the initial conditions are satisfied by y(x) =
− 15 2 15 15
2 y1 (x) = x − 2 x + 15 − 2x .
We can also attempt a power series solution around x = 1, say y(x) = ∞ n
P
n=0 an (x−1) , put
t = x−1, P then the differential equation becomes (t2 +2t+1)y 00 +(−2t2 +t+3)y 0 +(4t+7)y =
0, y(t) = ∞ n 2
n=0 an t . We get the recursion relation, (−2n + 6)an−1 + (n + 7)an + (2n +
2

5n + 3)an+1 + (n2 + 3n + 2)an+2 = 0, and we know from initial conditions a0 = 1,


a1 = 2. We obtain a2 = −13/2, a3 = 15/2, a4 = −15/2, a5 = 15/2. Note that putting
an = an+2 = 1,an−1 = an+1 = −1 the recusrion relation is satisfied. So in fact for k > 1,
15
a2k+1 = 15/2 and a2k = −15/2. Hence, one obtains y(t) = − 2(1+t) +15− 15 2 (t+1)+(t+1) .
2

Although the power series method is valid for 0 < x < 2, the solution is valid for x > 0.
By uniqueness theorem for linear ODEs, this is the unique solution for x > 0.

S-ar putea să vă placă și