You are on page 1of 8

UPSC Civil Services Main 1989 - Mathematics

Calculus
Sunder Lal
Retired Professor of Mathematics
Panjab University
Chandigarh

January 14, 2010

Question 1(a) If f is at least thrice continuously differentiable, then show that

h2 00
f (a + h) = f (a) + hf 0 (a) + f (a + h)
2!
where lies between 0 and 1, and prove that limh0 = 13 .

Solution. Let
(a + h x)2
(x) = f (a + h) f (x) (a + h x)f 0 (x) A
2!
where A is so determined that (a + h) = (a). Clearly (a + h) = 0, and (a) = f (a + h)
2 2
f (a) hf 0 (a) h2! A, so (a) = 0 h2! A = f (a + h) f (a) hf 0 (a).
Now satisfies the requirements of Rolles theorem in [a, a + h], therefore there exists
(0, 1) such that 0 (a + h) = 0, note that a + h (a, a + h). But

0 (a + h) = (h h)f 00 (a + h) + (h h)A = 0 A = f 00 (a + h)

Now from (a) = 0, we get

h2 00
f (a + h) = f (a) + hf 0 (a) + f (a + h)
2!
as required.
Now consider the equality
h2 00
f (a + h) f (a) hf 0 (a) 2
f (a) h2 f 00 (a + h) f 00 (a)
=
h3 2 h3

1
Taking the limit as h 0 on both sides, we get
2
f (a + h) f (a) hf 0 (a) h2 f 00 (a)
LHS = lim
h0 h3
f (a + h) f (a) hf 00 (a)
0 0
= lim LHospitals rule
h0 3h2
f 00 (a + h) f 00 (a)
= lim LHospitals rule
h0 6h
1 000
= f (a)
6
f 00 (a + h) f 00 (a) f 00 (a + h) f 00 (a)
RHS = lim = lim
h0 2h h0 2 h
000
= f (a)
2
Thus = 31 as h 0, provided f 000 (a) 6= 0.
The question above is a particular case of the following:
Taylors Theorem: Let f (n1) , the (n 1)-th derivative of a real valued function be
continuous in the closed interval [a, a + h], and let f (n) exist in the open interval (a, a + h),
then there exists a real number , 0 < < 1, such that
hn1 (n1)
f (a + h) = f (a) + hf 0 (a) + . . . + f (a) + Rn
(n 1)!
where
hn (1 )np (n)
Rn = f (a + h), 0 < < 1
(n 1)!p
Proof: The condition that f (n1) (x) is continuous in [a, a + h] f, f 0 , f 00 , . . . , f (n2) are
continuous in [a, a + h]. Let
(a + h x)n1 (n1)
(x) = f (x) + (a + h x)f 0 (x) + . . . + f (x) + A(a + h x)p
(n 1)!
where A is a constant to be determined so that (a + h) = (a) i.e.

0 (hn1 (n1)
f (a + h) = f (a) + hf (a) + . . . + f (a) + Ahp
(n 1)!
Now (i) (x) is continuous in [a, a + h] (ii) (x) is differentiable in (a, a + h) and (iii)
(a + h) = (a). Thus all requirements of Rolles theorem are satisfied,so there exists a real
number , 0 < < 1 such that 0 (a + h) = 0 (note that any real number c (a, a + h) can
be written as c = a + ca
h
h, and since a < c < a + h, we get 0 < = ca
h
< 1). But

(a + h x)n1 (n)
0 (x) = f (x) pA(a + h x)p1
(n 1)!

2
so
(h(1 ))n1 (n) hnp (1 )np (n)
0 (a+h) = f (a+h)pAhp1 (1)p1 = 0 A = f (a+h)
(n 1)! (n 1)!p

Thus
hn1 (n1) hn (1 )np (n)
f (a + h) = f (a) + hf 0 (a) + . . . + f (a) + f (a + h)
(n 1)! (n 1)!p

If we put n = p we get

hn1 (n1) hn
f (a + h) = f (a) + hf 0 (a) + . . . + f (a) + f (n) (a + h)
(n 1)! n!

The given question is for n = 2, so

h2 00
f (a + h) = f (a) + hf 0 (a) + f (a + h)
2!
Now we prove the second part: If f (n+1) (x) is continuous from the right at x = a, then
1
n+1 as h 0 provided f (n+1) (a) 6= 0.
In the given question, n = 2, and f (3) (x) is continuous in [a, a + h] so 13 .
f (n+1) (x) is continuous from the right at x = a there exists k, 0 < k < h, such that
f (n+1) (x) exists in [a, a + k]. Thus by the above theorem:

0 (k n1 (n1) k n (n)
f (a + k) = f (a) + kf (a) + . . . + f (a) + f (a + k)
(n 1)! n!
n n+1
(k (n) k
f (a + k) = f (a) + kf 0 (a) + . . . + f (a) + f (n+1) (a + k)
n! (n + 1)!

where 0 < < 1, 0 < < 1. Subtracting the first from the second,
k
f (n) (a + k) f (n) (a) = f (n+1) (a + k)
n+1

Now we use Lagranges mean value theorem for f (n) (x) in [a, a + k] and obtain

k
kf (n+1) (a + 1 k) = f (n+1) (a + k), 0 < 1 < 1
n+1
1
Taking limit as h 0 i.e. k 0, we get limh0 = n+1 as f (n+1) (a+1 k) and f (n+1) (a+k)
both tend to f (n+1) (a) as k 0, and f (n+1) (a) 6= 0.

Question 1(b) Prove that thevolume of a right circular cylinder of greatest volume which
can be inscribed in a sphere is 33 times that of a sphere.

3
Solution. Let the radius of the sphere be a, and the radius and height of the cylinder be r
and h respectively. Then r2 + ( h2 )2 = a2 . Letting r = a cos , h = 2a sin , the volume of the
cylinder is 2a3 cos2 sin .
dV
For extreme values, = 2a3 [cos3 2 cos sin2 ] = 0. Now cos = 0 = 2
d
r = 0, which is not admissible. So cos2 2 sin2 = 0 tan = 12 as 0 < < 2 . Thus
q
sin = 13 , cos = 23 .
d2 V 3
q
2
= 2a [3 cos sin + 2 sin 4 cos sin ] < 0 for sin = 3 , cos = 23 . Thus
3 2 2 1
d
V is maximum when tan = 12 .

V = 2a3 23 1 = 4 a3 3
3 3 3
3
Thus V is 3
times the volume of the sphere.

2z
Question 1(c) Show that at the point of the surface xx y y z z = c where x = y = z, =
x y
[x log ex]1 .

Solution. Taking logs on both sides of xx y y z z = c, we get

z log z + y log y + x log x = log c

Differentiating with respect to x,


z z 1 + log x log ex
(1 + log z) + (1 + log x) = 0 = = ()
x x 1 + log z log ez
z log ey
Similarly, = .
y log ez
Differentiating (*) w.r.t. y,

2z 1 z z 2z 1 z z log ex log ey
(1 + log z) + =0 = =
x y z y x x y z log ez y x z(log ez)3

2z log ex log ex
Letting x = y = z, we get = = [x log ex]1 as required.
x y x(log ex)3

Question 2(a) Find the surface of the solid generated by revolving x = a cos3 t, y = a sin3 t
about the x-axis.

4
Solution. We confine ourselves to the first quadrant because of symmetry. The curve is
2 2 2
x3 + y 3 = a3 .
Z
2 ds
S = 2 2y dt
0 dt
Z r 
2
3 dx 2  dy 2
= 4 a sin t + dt
0 dt dt
Z
2 1
sin3 t (3a cos2 t sin t)2 + (3a sin2 t cos t)2 2 dt

= 4a
0
Z
2
= 4a 3a sin4 t cos t dt
0
5
 2
2 sin t 12 2
= 12a = a
5 0 5

Question 2(b) If for a curve x sin + y cos = f 0 () and x cos y sin = f 00 () then show
that S = f () + f 00 () + C.

Solution. Solving for x and y, we get

x = f 0 () sin + f 00 () cos
y = f 0 () cos f 00 () sin
dx
= f 0 () cos + f 00 () sin f 00 () sin + f 000 () cos
d
= f 0 () + f 000 () cos


dy
= f 0 () sin + f 00 () cos f 00 () cos f 000 () sin
d
= f 0 () + f 000 () sin

Z r 2  2 Z
dx dy  0
f () + f 000 () d

S = + d =
d d
00
= f () + f () + C

as required.
ZZZ
1
Question 2(c) Evaluate (1 z) 2 dx dy dz over the interior of the tetrahedron with
faces x = 0, y = 0, z = 0, x + y + z = 1.

5
Solution.
Z 1 Z 1z Z 1yz
1
I = (1 z) 2 dx dy dz
z=0 y=0 x=0
Z1 Z1z
1
= (1 z) 2 (1 y z) dy dz
z=0 y=0
1
y2
Z h i1z
1
= (1 z) 2 y zy dz
z=0 2 0
Z 1
1
h (1 z)2 i
= (1 z) 1 z
2 z(1 z) dz
z=0 2
Z 1
3
h (1 z) i
= (1 z) 2 1 z dz
z=0 2
Z 1 1
1 5 1 2 7 1
= (1 z) 2 dz = (1 z) 2 =
z=0 2 2 7 0 7

Paper II

Question 3(a) Show that the value of


ZZ
1 1
(1 x y)3 x 2 y 2 dx dy

taken over the interior of the triangle whose vertices are the origin and the points (0, 1) and

(1, 0) is 480 .

Solution. We will convert this into a Dirichlet integral. Let x + y = u, y = uv, so


(x, y) 1 v u
= u. u varies from 0 to 1, and v = y = y also
x = u(1 v), and = u x+y
(u, v) v u
varies from 0 to 1.

Z 1 Z 1
1 1 1 1
I = (1 u)3 u 2 (1 v) 2 u 2 v 2 u du dv
Z0 1 Z0 1
3 3
= (1 u)41 u31 (1 v) 2 1 v 2 1 du dv
0 0
(4)(3) ( 23 )( 32 )
=
(7) (3)
3!
= =
6! 4 480
as ( 12 ).

6
Question 3(b) Obtain the largest and the least values of 2(x + y + z) xyz on the closed
ball x2 + y 2 + z 2 9.

Solution. Let F (x, y, z) = 2(x + y + z) xyz (x2 + y 2 + z 2 9), where is Lagranges


undetermined multiplier. For extreme values:
F F F
= 2 yz + 2x = 0, = 2 xz + 2y = 0, = 2 yx + 2z = 0
x y z
Subtracting the first two, z(x y) + 2(x y) = 0 x = y or z = 2. Similarly from the
other pairs, we get y = z or x = 2 and x = z or y = 2.
We now explicitly find extreme vaues to get the greatest and least values of f (x, y, z) =
2(x + y + z) xyz

1. x = y = z: This gives us x2 + x2 + x2 = 9, so x = 3.

(a) x = y = z = 3: f (x, y, z) = 6 3 3 3 = 3 3.

(b) x = y = z = 3: f (x, y, z) = 6 3 + 3 3 = 3 3.

2. x = y, y = 2: 42 + 42 + z 2 = 9 z 2 = 9 82 . Using 2 xy + 2z = 0, we have
22 1
2 42 + 2z = 0 z = . Thus

 22 1 2
= 9 82

44 42 + 1 = 92 84
124 132 + 1 = 0
1
2 = 1,
12
= 1 x = y = 2, z 2 = 1
= 1 x = y = 2, z 2 = 1
1 1 25
= x = y = , z2 =
2 3 3 3
1 1 25
= x = y = , z2 =
2 3 3 3
We examine each of these 8 cases:

(a) x = y = 2, z = 1: f (x, y, z) = 6 4 = 10.


(b) x = y = 2, z = 1: f (x, y, z) = 10 + 4 = 6.
(c) x = y = 2, z = 1: f (x, y, z) = 10 4 = 6.
(d) x = y = 2, z = 1: f (x, y, z) = 6 + 4 = 10.
(e) x = y = 13 , z = 5 :
3
f (x, y, z) = 6
3
5

3 3
= 13
.
3 3

7
(f) x = y = 13 , z = 53 : f (x, y, z) = 143 + 5

3 3
= 337
.
3

(g) x = y = 1 , z = 5 : f (x, y, z) = 14
5
= 37
.
3 3 3 3 3 3 3

(h) x = y = 1 , z = 53 : f (x, y, z) = 63 + 5
= 313
.
3 3 3 3

We need not consider the other possibilities like y = z, z = 2 as the situation is symmetric
and no new values will result. The greatest value of f (x, y, z) is 10, and the least value is
-10.
We dont consider d2 F , as it is not needed. We use the fact the all the maximum and
minimum values occur at the extreme values. Calculation of d2 F is a very lengthy process.

You might also like