You are on page 1of 7

UPSC Civil Services Main 1983 - Mathematics

Calculus
Sunder Lal
Retired Professor of Mathematics
Panjab University
Chandigarh

January 16, 2010

Question 1(a) Consider the function

log(2 + x) x2n sin x


f (x) = lim
n 1 + x2n
in the interval 0 x 2 . Show that the function does not vanish anywhere in this interval
and explain why is so although f (0), f ( 2 ) differ in sign.

Solution. Since x2n 0 as n for 0 x < 1, it follows that


log 3 sin 1
f (x) = log(2 + x), 0 x < 1, f (1) =
2
1 log(2+x)
For x > 1, x2n
0 and x2n
0 as n , therefore
log(2+x)
x2n
sin x
f (x) = lim 1 = sin x
n
x2n
+1

Thus
log(2 + x), 0 x < 1

f (x) = log 3sin
2
1)
, x=1

sin x, 1 < x 2

Clearly f (0) = log 2 > 0, f ( 2 ) = sin 2 = 1 < 0, and f (x) 6= 0 for any x [0, 2 ].
The reason why this can happen is that f (x) is not continuous in the interval [0, 2 ], so
the intermediate value theorem does not apply. The function is discontinuous at x = 1 as
limx1+ f (x) = sin 1, limx1 f (x) = log 3 and both of these are different from f (1).

1
Question 1(b) Prove that there is a point a (h, h) such that
Z h
2h3 00
f (x) dx = h[f (h) + f (h)] f (a)
h 3
Solution. Integrating by parts, we get
Z h ih Z h Z h
0
f (x) dx = xf (x) xf (x) dx = h[f (h) + f (h)] xf 0 (x) dx
h h h h
R0
Now put x = t in h
xf 0 (x) dx to get
Z 0 Z 0 Z h
0 0
xf (x) dx = (t)f (t)(dt) = tf 0 (t) dt
h h 0

Thus Z h Z h
x f 0 (x) f 0 (x) dx

f (x) dx = h[f (h) + f (h)]
h 0
0 0
f (x) f (x) f (x) + f 00 (x)
00
f 0 (x) f 0 (x)
We now note that lim = lim = 2f 00 (0), showing that
x0 x x0 1 x
is Riemann integrable in [0, h]. Since x2 keeps the same sign, the first mean value theorem
of integral calculus gives us
Z h 0 0
2 f (x) f (x) f 0 () f 0 () h 2
Z
x dx = x dx
0 x 0

for some (0, h). Consequently,


Z h
f 0 () f 0 () 2h3
f (x) dx = h[f (h) + f (h)]
h 2 3
We now apply Lagranges mean value theorem to the function f 0 (x) in [, ] and get
f 0 () f 0 ()
= f 00 (a) for some a (, )
2
Thus
h
2h3 00
Z
f (x) dx = h[f (h) + f (h)] f (a)
h 3
Note that 0 < < h a [h, h].
Corollary: If f 000 (x) exists in (h, h) prove that
f (h) f (h) f 0 (h) + f 0 (h) h3 000
=h f (a)
2 2 3
for some a (h, h).
Proof: Use the above result for f 0 (x) instead of f (x). The left hand side now becomes
f (h) f (h), and dividing by two, we get the above equation.

2
Question 1(c) Find the least perimeter of an isoceles triangle in which a circle of radius r
can be inscribed.

Solution.

Let be the semi-vertical angle. D is the A


midpoint of BC. E is the point of contact
of AC and the circle or radius r inscribed
OE
in ABC. Therefore = tan , or AE =
AE F E
OE cot = r cot . Similarly, we find out r r
that OA = r csc , and AD = AO + OD = O
r(1 + csc ), consequently BD = AD tan =
r(1 + csc ) tan . Since 4ABC is isoceles,
BD = DC, DC = CE and AE = AF . B D C

The perimeter

P = BD + DC + CE + AE + BF + AF
= 4BD + 2AE = 4r(1 + csc ) tan + 2r cot
= 4r tan + 4r sec + 2r cot
dP
= 4r sec2 + 4r sec tan 2r csc2 = 0
d
2 2 sin 1
0 = +
cos2 cos2 sin2
0 = 2 sin + 2 sin cos2
2 3

0 = 2 sin3 + 3 sin2 1
0 = (sin + 1)2 (2 sin 1)

Thus dP
d
= 0 sin = 21 , since sin = 1 is not possible as 2 is the angle of a triangle.
Thus = 6 .

d2 P
= 8r sec2 tan + 4r(sec tan2 + sec3 ) 4r csc ( csc cot ) > 0
d2
for = 6 . Therefore we have a minimum when = 6 . The required perimeter is 4r 13 +

4r 23 + 2r 3 = 6r 3.

Question 2(a) If f (0) = 0 and f 0 (x) = 1+x


1
2 , prove without using the method of integration

that  x+y 
f (x) + f (y) = f
1 + xy

3
x+y
Solution. Let u = f (x) + f (y), v = 1+xy
. Then

v (1 xy)(1) (x + y)(y) 1 + y2
= =
x (1 xy)2 (1 xy)2
v (1 xy)(1) (x + y)(x) 1 + x2
= =
y (1 xy)2 (1 xy)2

(u, v) 12 1
1+x 1+y 2
= 1+y2 1+x = 0
2
(x, y) (1xy)2 (1xy) 2

 x+y 
Therefore u, v are functionally dependent. Let u = (v) f (x) + f (y) = .
1 xy
Substituting y =  x0+gives us f (x) + f (0) = (x) or (x) = f (x) as f (0) = 0. Hence
y
f (x) + f (y) = f , which was to be proved.
1 xy

Question 2(b) Show that under the transformation u = x2 y 2 , v = 2xy the equation

2H 2
2 H H H
y2 2
x 2
=x y
x y x y
 
H
becomes u v = 0.
v u v

Solution.
H H u H v H H
= + = 2x + 2y
x u x v 2x u  v  2
2H 2
2H

H H H H
= 2 + 2x 2x + 2y + 2y 2x + (2y)
x2 u u2 u v v u v 2
H 2H 2H 2H
= 2 + 4x2 2 + 8xy + 4y 2 2 (1)
u u u v v
H H u H v H H
= + = 2y + 2x
y u y v y u v
2
 2 2
 2
2H
 
H H H H H
= 2 2y (2y) + 2x + 2x (2y) + (2x)
y 2 u u2 u v v u v 2
H 2H 2H 2H
= 2 + 4y 2 2 8xy + 4x2 2 (2)
u u u v v
H H H H H H
x y = 2x2 + 2xy + 2y 2 2xy
x y u v u v
H
= 2(x2 + y 2 ) (3)
u

4
Using (1), (2), (3) we get
2 2
 
2 H 2 H H H
0 = y x x y
x2 y 2 x y
2 2 2
2 H 2 2 H 3 H 4 H
= 2y + 4x y + 8xy + 4y
u u2 u v v 2
2 2
H H H 2H H
+2x2 4x2 y 2 2 + 8x3 y 4x4 2 2(x2 + y 2 )
u u u v v u
2 2
H H
= 8xy(x2 + y 2 ) + 4(y 4 x4 ) 2
u v v
2 2
H H
0 = (y 2 x2 ) 2 + 2xy
v u v
2H 2H
0 = u 2 + v
 v u
 v
H
0 = u v
v u v
as required.

Question 2(c) Prove that the volume of the solid generated by the revolution of the tractrix
t
x = a cos t + a log tan , y = a sin t
2
about the asymptote is equal to half the volume of a sphere of radius a.
Solution. The x-axis is the asymptote, as t 0 x , y 0. The volume required
is twice the volume in the first quadrant.

Z
2 dx
V = 2 y 2 dt
0 dt
Z
2 a
= 2 a2 sin2 t(a sin t + ) dt
0 sin t
Z
2
3
= 2a (sin t sin3 t) dt
0
Z
2
3
= 2a sin t cos2 t dt
0

cos3 t 2 2a3

3
= 2a =
3 0 3
3
1 4a 1
= = Volume of a sphere of radius a
2 3 2

5
Paper II

Question 3(a) Obtain a set of sufficient conditions such that for a function f (x, y)

2f 2f
=
x y y x

Solution. Eulers Theorem: If the partial derivatives fxy and fyx are continuous at (a, b)
then fxy (a, b) = fyx (a, b).
Proof: Let (h, k) = f (a + h, b + k) f (a + h, b) f (a, b + k) + f (a, b) where (a + h, b +
k), (a + h, b), (a, b + k) all belong to a neighborhood N of (a, b) we can take for N an open
disc with center (a, b) in which fxy and fyx exist.
Let G(x) = f (x, b + k) f (x, b) for x Ih where Ih = [a, a + h] or [a + h, a] according
as h > 0 or h < 0. Clearly G0 (x) = fx (x, b + k) fx (x, b) for x Ih . We apply Lagranges
Mean Value Theorem to G(x) and obtain:

G(a+h)G(a) = (h, k) = hG0 (a+h) = h[fx (a+h, b+k)fx (a+h, b)] ()

where 0 < < 1.


Now we consider F (t) = fx (a + h, t) for t Ik where Ik = [b, b + k] or [b + k, b] according
as k > 0 or k < 0. We apply Lagranges Mean Value Theorem to F (t) and obtain:

F (b + k) F (b) = fx (a + h, b + k) fx (a + h, b) = kF 0 (b + 1 k)

where 0 < 1 < 1. But F 0 (t) =


f (a
y x
+ h, t), so

2f
F (b + k) F (b) = k (a + h, b + 1 k)
y x

Using (*) we get


2f
(h, k) = hk (a + h, b + 1 k)
y x
Since fyx is continuous at (a, b), we get

(h, k) 2f 2f
lim = lim (a + h, b + 1 k) = (a, b)
h0,k0 hk h0,k0 y x y x

Now instead of G(x), we start with H(y) = f (a+h, y)f (a, y) for y Ik , and proceeding
exactly as above, we get
(h, k) 2f
lim = (a, b)
h0,k0 hk x y
2f 2f
Hence (a, b) = (a, b), which completes the proof.
x y y x

6
Question 3(b) Find the maximum and minimum values of x2 + y 2 + z 2 subject to the
conditions x + y + z = 1, xyz + 1 = 0.

Solution. Let F (x, y, z) = x2 + y 2 + z 2 + 1 (x + y + z 1) + 2 (xyz + 1) where 1 , 2 are


Lagranges undetermined multipliers. The extreme values are given by
F
= 2x + 1 + 2 yz = 0
x
F
= 2y + 1 + 2 xz = 0
y
F
= 2z + 1 + 2 xy = 0
z
Subtracting the first two, 2(x y) + 2 z(y x) = 0 x = y or 2 = z2 (Note that
x 6= 0, y 6= 0, z 6= 0 because xyz + 1 = 0). Similarly from the other pairs of equations, we
get y = z or 2 = x2 , and x = z or 2 = y2 .
Since xyz = 1, x + y + z = 1, it follows that x, y, z cannot be all positive or all negative,
moreover two must be positive and one negative. In particular, x = y = z is not possible.
Suppose x 6= y, then 2 = z2 . Now F x
= 0 2x + 1 + 2y = 0.
Substituting the values of 1 , 2 in F
z
= 0, we get

2xy
2z 2(x + y) + =0
z
But x + y = 1 z, xy = z1 , so we get 2z 2(1 z) z22 = 0 2z 3 z 2 1 = 0
(z 1)(2z 2 + z + 1) = 0. But 2z 2 + z + 1 has no real roots, so the only real root is z = 1.
Thus we get 2 = 2, z = 1, 1 = 0( x + y + z = 1 x + y = 0 1 = 0), x = 1, y =
1( x + y = 0, xy = 1 x2 = 1 x = 1, y = 1). Hence the stationary values are
(1, 1, 1), (1, 1, 1) and f (x, y, z) = 3 at these points.
On taking y 6= z we shall get (1, 1, 1), (1, 1, 1) as stationary points (by symmetry),
and for x 6= z, we get (1, 1, 1), (1, 1, 1). f (x, y, z) is always equal to 3, hence we cannot
say whether it is a maximum or minimum without checking d2 F .
Considering the point (1, 1, 1),

d2 F = 2(dx)2 + 2(dy)2 + 2(dz)2 + 4 dx dy + 4 dx dz 4 dy dz

Now x+y+z = 1 dx+dy+dz = 0 dz = dxdy. xyz = 1 yz dx+zx dy+xy dz =


0 dx dy dz = 0 dx = 0, dz = dy.
Thus d2 F = 8(dz)2 > 0, so f has a minimum at all these stationary points.
Note: The question can be treated as that of one variable as y, z can be eliminated, but
the calculation becomes quite messy.

You might also like