UPSC Civil Services Main 1979 - Mathematics

Calculus
Sunder Lal
Retired Professor of Mathematics
Panjab University
Chandigarh

January 16, 2010

Question 1(a) If f (x) C n+1 for |x a| h, then show that

n Z x (n+1)
X f (k) (a) k f (t)(x t)n
f (x) = (x a) + dt
k=0
k! a n!

Further show that the remainder after n + 1 terms can be expressed as

f (n+1) (X)(x X)n
(x a)
n!
where a < X < x.
Solution. Carrying out integration by parts on the given integral repeatedly, we get
Z x (n+1) x Z x (n)
f (t)(x t)n f (n) (t)(x t)n f (t)(x t)n1
dt = + dt
a n! n! a a (n 1)!
Z x (n)
f (n) (a)(x a)n f (t)(x t)n1
= + dt
n! a (n 1)!
Z x (n1)
f (n) (a)(x a)n f (n1) (a)(x a)n1 f (t)(x t)n2
= + dt
n! (n 1)! a (n 2)!
= ...
n Z x
X f (r) (a)(x a)r
= + f 0 (t) dt
r=1
r! a
n
X f (r) (a)(x a)r
= + f (x) f (a)
r=1
r!
n
X f (r) (a)(x a)r
= + f (x)
r=0
r!

1
Thus n x
f (r) (a)(x a)r f (n+1) (t)(x t)n
X Z
f (x) = + dt
r=0
r! a n!
as required. Now we use the result if f (x) is continuous in [a, b], then
Z b
f (x) dx = f ()(b a) for some (a, b)
a
Rb
This is obvious as m(b a) a f (x) dx M (b a), where m, M are respectively the
min and max of f (x) over [a, b]. and every value between m and M is attained (by the
intermediate value theorem).
Since f C n+1 , f (n+1) is continuous and we can use the above result. Therefore
Z x (n+1)
f (t)(x t)n f (n+1) (X)(x X)n
dt = (x a)
a n! n!
where a < X < x.

Question 1(b) If  2 2

2xy x y , (x, y) 6= (0, 0)

f (x, y) = x2 + y 2
0, (x, y) = (0, 0)

2f 2f
show that 6= . Explain this result.
x y y x
Solution.
f (0, k) f (0, 0) 0
fy (0, 0) = lim = lim = 0
k0 k k0 k
2 k 2
f (h, k) f (h, 0) 2hk hh2 +k 2
fy (h, 0) = lim = lim = 2h
k0 k k0 k
2f fy (h, 0) fy (0, 0) 2h
(0, 0) = lim = lim =2
x y h0 h h0 h

f (h, 0) f (0, 0) 0
fx (0, 0) = lim = lim = 0
h0 h h0 h
2 2
f (h, k) f (0, k) 2hk hh2 k
+k2
fx (0, k) = lim = lim = 2k
h0 h h0 h
2f fx (0, k) fx (0, 0) 2k
(0, 0) = lim = lim = 2
y x k0 k k0 k

2f 2f
Thus 6= at (0, 0). The reason for this is that neither fxy nor fyx is continuous
x y y x
at (0, 0). This implies that neither fx (x, y) nor fy (x, y) is differentiable at (0, 0). Thus the
criteria of Youngs or Schwartz theorem are not satisfied.

2
 R1
Question 2(a) If B(x, y) = 0
tx1 (1 t)y1 dt, x > 0, y > 0, then show that B(x, x) =
212x B(x, 21 ).

Solution. See 1982, question 2(b), where we proved

( 12 )(x) 22x1 [(x)]2

=
(x + 12 ) (2x)

( 21 )(x) 1 [(x)]2
Since = B(x, ) and = B(x, x), we get B(x, x) = 212x B(x, 12 ) as required.
(x + 12 ) 2
(2x)

Paper II

Question 3(a) Find the maximum of x21 x22 . . . x2n under the restriction x21 +x22 +. . .+x2n = 1.
Using the result derive the inequality
1 a1 + . . . + an
(a1 a2 . . . an ) n
n
for positive real numbers a1 , . . . , an .

Solution. Let F (x1 , . . . , xn ) = x21 x22 . . . x2n + ( ni=1 x2i 1), where is Lagranges unde-
P
termined multiplier. For extreme values,
F
= 2x21 x22 . . . x2i1 xi x2i+1 . . . x2n + 2xi = 0, 1 i n
xi
Since x21 +x22 +. . .+x2n = 1, (x1 , . . . , xn ) 6= (0, . . . , 0), it follows that = x21 x22 . . . x2i1 x2i+1 . . . x2n
for 1 i n. Thus x21 = x22 = . . . = x2n = n1 , and = nn1 1
.

2F
= 2x21 x22 . . . x2i1 x2i+1 . . . x2n + 2 = 0 for 1 i n
x2i
n
2F Y 4 1
= 4xi xj x2r = at x21 = x22 = . . . = x2n =
xi xj r=1,r6=i,j
nn1 n
8 X
d2 F = dxi dxj (only the cross terms appear)
nn1 1i<jn

Now x21 + x22 + . . . + x2n = 1, so 2 ni=1 xi dxi = 0 ni=1 dxi = 0 dxn = (dx1 + . . . +
P P
dxn1 ).
n1
2 8 h X X 2 i
Thus d F = n1 dxi dxj dxi . Clearly d2 F < 0, so we have a
n 1i<jn1 i=1
maximum when x21 = x22 = . . . = x2n = n1 .

3
 Derivation of inequality:
 1 n  x2 + x2 + . . . + x2  n
1 2 n
x21 x22 . . . x2n = as x21 + x22 + . . . + x2n = 1
n n
2 2 2
1 x + x2 + . . . + xn
(x21 x22 . . . x2n ) n 1
n
Let x2i = ai , 1 i n, to get
1 a1 + . . . + an
(a1 a2 . . . an ) n
n
as required.
x 2 1 2 2
ex (t +1)
Z Z
t2
Question 4(a) Define f (x) = e dt , g(x) = 2
dt. Show that g 0 (x) +
0 0 tZ + 1
x
t2
f 0 (x) = 0 for all x. Deduce that (i) g(x) + f (x) =
4
(ii) lim e dt = .
x 0 2
Solution.
Z x 
0 t2
2
f (x) = 2 e dt ex
0
1 2 2
ex (t +1)
Z  
0
g (x) = dt
0 x t2 + 1
1 2 (t2 +1)
ex (2x(t2 + 1))
Z
= dt
0 t2 + 1
2 Rx 2
Let xt = y in the second equation, then dy = x dt, g 0 (x) = 2ex 0
ey dy = f 0 (x),
showing that f 0 (x) + g 0 (x) = 0 for all x.
R1
1. f 0 (x) + g 0 (x) = 0 f (x) + g(x) = C, a constant. Clearly f (0) = 0, g(0) = dt
0 t2 +1
=
i1
tan1 t = 4 . So C = f (0) + g(0) = 4 , thus g(x) + f (x) = 4 .
0
2 2
1
ex t
Z 
2
2. lim g(x) = lim 2
dt ex .
x x 0 t +1
Z 1 x2 t2 Z 1
e dt 2 2 2
But 2
dt 2
= , ex t 1. Thus 0 g(x) 4 ex
0 t +1 0 1+t 4
limx g(x) = 0.
Z x 2 Z x
t2 t2
Thus lim f (x) = lim e dt = = lim e dt = , as both sides are
x x 0 4 x 0 2
positive.