MADHAVA MATHEMATICS COMPETITION, 12th January 2020

Solutions and scheme of marking

N.B.: Part I carries 20 marks, Part II carries 30 marks and Part III carries
50 marks.
Part I
N.B. Each question in Part I carries 2 marks.

1. Let A be a non-empty subset of real numbers and f : A → A be a function such

that f (f (x)) = x for all x ∈ A. Then f (x) is
A) a bijection B) one-one but not onto
C) onto but not one-one D) neither one-one nor onto.
Answer: A
If f (x) = f (y), then f (f (x)) = f (f (y)) implies x = y. Therefore f is one-one
function. By definition, f is onto. Hence f is a bijection.

2. If f : R → R be a function satisfying f (x + y) = f (xy) for all x, y ∈ R and

f (3/4) = 3/4, then f (9/16) =
A) 9/16 B) 0 C) 3/2 D) 3/4.
Answer: D
If we put y = 0, then f (x) = f (0) for all x ∈ R. This implies that f is a constant
function.

3. The area enclosed between the curves y = sin2 x and y = cos2 x in the interval
0 ≤ x ≤ π/2 is
A) 2 B) 1/2 C) 1 D) 3/4.
Answer: C
Z π/4 Z π/2
Area = (cos2 x − sin2 x)dx + (sin2 x − cos2 x)dx
Z π/4 0 Z π/2 π/4

= cos 2xdx − cos 2xdx = 1/2 + 1/2 = 1.

0 π/4

m 12
4. The number of ordered pairs (m, n) of all integers satisfying = is
12 n
A) 15 B) 30 C) 12 D) 10.
Answer: B
We have mn = 144 and 144 = 24 32 . There are 15 divisors of 144 which are
positive integers. Including negative integers total 30 pairs are there.

5. Suppose 2 log x + log y = x − y. Then the equation of the tangent line to the
graph of this equation at the point (1, 1) is
A) x + 2y = 3 B) x − 2y = 3 C) 2x + y = 3 D) 2x − y = 3.
Answer: A
Consider 2 log x + log y = x − y. Differentiating this equation with respect to x,

1
 2 1 dy dy
we get + = 1− . At point (1, 1) the slope of tangent is −1/2. The
x y dx dx
equation of tangent is x + 2y = 3.

6. Let f : R → R defined as f (x) = sin[x], where [x] denotes the greatest integer
less than or equal to x. Then
A) f is a 2π-periodic function B) f is a π-periodic function
C) f is a 1-periodic function D) f is not a periodic function.
Answer: D
If f is a periodic function with period T, then sin[x + T ] = sin[x]. This implies
[x + T ] − [x] = 2nπ for some integer n, which is not possible.

7. For how many integers a with 1 ≤ a ≤ 100, aa is a square?

A) 50 B) 51 C) 55 D) 56.
Answer: C
Case 1: all even integers. There are 50 even integers between 1 and 100.
Case 2: a is an odd number which is a square i. e. 1,9,25,49,81.
Therefore total number is 55.
 
1
8. lim x
x→0 x
A) 0 B) 1 C) −1 D) does not exist.
Answer:
  B        
1 1 1 1 1
lim x − 1 = lim x −x = lim x − = 0.
x→0 x x→0 x x x→0 x x
 2  2
2 α β
9. If α and β are the roots of x + 3x + 1 then + equals
β+1 α+1
A) 19 B) 18 C) 20 D) 17.
Answer: B
Observe that α + β = −3, αβ = 1, α2 + 3α + 1 = 0, β 2 + 3β + 1 = 0. Therefore
(α + 1)2 = α2 + 2α + 1 = −α, (β + 1)2 = β 2 + 2β + 1 = −β.
 2  2  2   2 
α β α β 1 + 3α 1 + 3β
+ = + = +
β+1 α+1 −β −α β α
(1 + 3α)α + (1 + 3β)β
= = 3(α2 + β 2 ) + (α + β) = 18.
αβ
10. The equation z 3 + iz − 1 = 0 has
A) no real root B) exactly one real root
C) three real roots D) exactly two real roots.
Answer: A
If x is a real root, then x3 + ix − 1 = 0. This implies that x = 0 and x3 = 1,
which is not possible.

2
 Part II
N.B. Each question in Part II carries 6 marks.

1. Let a1 , a2 , · · · be a sequence of natural numbers. Let (a, b) denote the greatest

common divisor (gcd) of a and b. If (am , an ) = (m, n) for all m 6= n, prove that
an = n for all n ∈ N.
Solution: Note that (an , an2 ) = (n, n2 ) = n. Therefore n divides an . [2]
Let an = mn. Then (an , amn ) = (n, mn) = n and (am , amn ) = (m, mn) = m. [2]
This implies that mn divides amn . Since mn divides both amn and an , it divides
their gcd n. Hence m = 1 and thus an = n. [2]

2. Let f : C → C be a function such that f (z)f (iz) = z 2 for all z ∈ C. Prove that
f (z) + f (−z) = 0 for all z ∈ C. Find such a function.
Solution: It is given that f (z)f (iz) = z 2 for all z ∈ C. Replacing z by iz, we get
f (iz)f (−z) = −z 2 . Adding these two expressions we get, f (iz)[f (z)+f (−z)] = 0.
From f (z)f (iz) = z 2 we deduce that f (z) = 0 if and only if z = 0.
If z 6= 0, then f (iz) 6= 0 and so f (z) + f (−z) = 0.
If z = 0, then f (z)+ f (−z) =  2f (0) = 0. Thus  = 0, ∀z ∈ C. [4]
 f (z) + f (−z)
−1 1 1 1
Example: f (z) = √ + i √ z or f (z) = √ − i √ z [2]
2 2 2 2
3. Let n be a positive integer. Line segments can be drawn parallel to edges of a
given rectangle. What is the minimum number of line segments (not necessarily of
same lengths) that are required so as to divide the rectangle into n subrectangles?
Justify.

For example, in the adjacent figure, 3 segments are drawn to get 5

subrectangles and 3 is the minimum number.
Solution: It can be observed that there are three cases.
Case 1: Let n = k 2 . Consider (k−1) horizontal and (k−1) vertical line segments.
These 2(k − 1) line segments will generate k 2 subrectangles. Note that this is a
configuration with minimum number of line segments to divide the rectangle into
k 2 subrectangles. [2]
2 2
Case 2: Let k + 1 ≤ n ≤ k + k. Consider a small horizontal line segment,
which divides one of k 2 subrectangles into 2 subrectangles resulting into total
k 2 + 1 subrectangles. Extending this small line segment, we can get up to k 2 + k
subrectangles. Hence, in this case the the minimum number of line segments
required to divide the rectangle into n subrectangles is 2(k − 1) + 1 = 2k − 1. [3]
Case 3: Let k 2 + k + 1 ≤ n ≤ k 2 + 2k + 1. Applying the same procedure as given
in case 2, but instead of horizontal line segment we need to take a vertical line
segment and extend it as above. Thus, in this case the minimum number of line
segments required to divide the rectangle into n subrectangles is (2k−1)+1 = 2k.
[1]

3
 1 Z
1
4. Let f : [0, 1] → (0, ∞) be a continuous function satisfying f (t)dt = . Show
Z c 0 3
1
that there exists c ∈ (0, 1) such that f (t)dt = c − .
Z x 0 2
1
Solution: Define g(x) = f (t)dt + . [2]
0 2
Then g : [0, 1] → [0, 1] is a continuous function. By fixed point theorem,
Z c exists c ∈ (0, 1) such that g(c) = c. Observe that c 6= 0, 1. Thus we get
there
1
f (t)dt = c − . [4]
0 2
 
−1 1
5. Let A = . Show that there exist matrices X, Y such that A = X 3 +Y 3 .
0 −2
Solution: Note that A2 + 3A + 2I = 0. [2]
2 3 2
Therefore A(A + 3A + 2I) = A + 3A + 2A = 0. This implies that
(A + I)3 = A3 + 3A2 + 3A + I = A + I. [2]
Hence A = (A + I)3 − I =  (A + I)3+ (−I)3 .  
0 1 −1 0
Thus we get X = A + I = and Y = −I = . [2]
0 −1 0 −1
Part III

 f : (0,∞) → R be a continuous function satisfying f (1) = 5 and

1. Let
x
f = f (x) + 2 for all positive real numbers x.
x+1
a) Find lim f (x).
x→∞
b) Show that lim+ f (x) = ∞.
x→0
c) Find one example of such a function. [12]
x
Solution: a) Note that as x → ∞, → 1.
x+1
Hence lim f (x) = f (1) − 2 = 5 − 2 = 3. [2]
x→∞
x
b) Observe that as x → 0+ , → 0. If limx→0+ f (x) = L, then L = L + 2.
x+1
This implies that the limit is not finite. [1]
1 1 1 1 x 1 1
Define g : [ n+1 , n ] → [ n+2 , n+1 ] as g(x) = . Note that g( n+1 ) = n+2 ,
x+1
g( n1 ) = n+1
1
and g 0 (x) = (x+1)1
2 > 0. Therefore g is an increasing function and

hence g is one-one and onto. For x ∈ [ 13 , 12 ], there is some t ∈ [ 12 , 1] such that

t
f (x) = f (g(t)) = f ( t+1 ) = f (t) + 2. By induction, it can be proved that for
x ∈ [ n+1 , n ], there is some t ∈ [ 12 , 1] such that f (x) = f (t) + 2(n − 1).
1 1

Let M > 0 be any real number. Since f is continuous on [ 12 , 1], it is bounded and
attains its bounds. There exists M1 > 0 such that −M1 < f (t) < M1 , ∀t ∈ [ 12 , 1].
Suppose f (t0 ) = min f (t), t ∈ [ 12 , 1]. Therefore f (t) ≥ f (t0 ), ∀t ∈ [ 21 , 1]. Choose
n1 ∈ N such that f (t0 ) + 2(n1 − 1) > 0. We can choose n2 ∈ N such that

4
 n2 > M + M1 + 2. Suppose n0 = max{n1 , n2 } and δ = n10 > 0. If 0 < x < δ, then
x ∈ [ n01+1 , n10 ] and |f (x)| = |f (t) + 2(n0 − 1)|, t ∈ [ 21 , 1].
Now f (t) + 2(n0 − 1) ≥ f (t0 ) + 2(n0 − 1) ≥ f (t0 ) + 2(n1 − 1) > 0.
Note that |f (x)| = f (t) + 2(n0 − 1) ≥ −M1 + 2(n0 − 1) > −M1 − 2 + n0 ≥
−M1 − 2 + n2 > M.
Hence lim+ f (x) = ∞. [6]
x→0
2
c) Example: f (x) = 3 + . [3]
x
2. An n × n matrix A = (aij ) is given. The sum of any n entries of A, whose any
two entries lie on different rows and different columns, is the same.
a) Prove that there exist numbers x1 , x2 , · · · , xn and y1 , y2 , · · · , yn such that
aij = xi + yj for all i, j, 1 ≤ i, j ≤ n.
b) Prove that rank(A) ≤ 2. [12]
Solution: a) Consider n entries situated on different rows and different columns
aiji , i = 1, 2, · · · , n. Fix k and l, 1 ≤ k < l ≤ n and replace akjk and aljl with akjl
and aljk respectively. The new n entries are still situated on different rows and
different columns. Since the sums of sets of n entries are equal, this implies
akjk + aljl = akjl + aljk . [*]
Now denote by x1 , x2 , · · · , xn the entries in the first column and by x1 , x1 +y2 , x1 +
y3 , · · · , x1 + yn the entries in the first row.
 
x1 x1 + y2 x1 + y3 · · · x1 + yn
 x2 ··· 
A =  ..
 

. 
xn ···

That is we have defined xk = ak1 for all k, y1 = 0 and yk = a1k − ak1 for all k ≥ 2.
Now aij = xi + yj for all i, j with i = 1 or j = 1. Consider i, j > 1. From [*]
we deduce that a11 + aij = a1j + ai1 . Hence x1 + aij = xi + x1 + yj . This implies
aij = xi + yj for all i, j. [8]
th
b) LetAj denote the
   j column of
  matrix A. Then    
x1 + y j x1 1 x1 1
 x2 + y j   x2  1  x2  1
Aj =  ..  =  ..  +yj  ..  = α+yj β, where α =  ..  and β =  ..  .
         
 .  . . . .
xn + y j xn 1 xn 1
That is Aj ∈ hα, βi , where hα, βi denotes the linear span of α, β. This is true for
every column of A. Hence rank(A) ≤ dim hα, βi ≤ 2. [4]

3. Let I ⊆ R be an interval and f : I → R be a differentiable function. Let

 
f (b) − f (a)
J= : a, b ∈ I, a < b .
b−a

5
 Show that a) J is an interval.
b) J ⊆ f 0 (I) and f 0 (I) − J contains at most two elements. [13]
Solution: a) Note that J is an interval if and only if for every a, b ∈ J, a < b, we
f (b1 ) − f (a1 ) f (b2 ) − f (a2 )
have (a, b) ⊂ J. Let u, v ∈ J, u < v. Hence u = ,v = .
b 1 − a1 b2 − a2
Suppose p ∈ (u, v). Define a function Q on [0, 1] as

f (tb1 + (1 − t)b2 ) − f (ta1 + (1 − t)a2 )

Q(t) = .
t(b1 − a1 ) + (1 − t)(b2 − a2 )

Since Q is a continuous function and Q(0) = v, Q(1) = u, we deduce that there

f (b0 ) − f (a0 )
exists t0 such that Q(t0 ) = p. Hence p = , where b0 = t0 b1 +(1−t0 )b2
b 0 − a0
and a0 = t0 a1 + (1 − t0 )a2 . Thus p ∈ J. [6]
f (b) − f (a)
b) Using Lagrange’s Mean Value Theorem, we get = f 0 (c) for some
b−a
c ∈ (a, b). Hence J ⊆ f 0 (I). [2]
f (p + (1/n)) − f (p)
Let p ∈ I and yn = . If p is the right end point of an interval
1/n
f (p) − f (p − (1/n))
I, then choose yn = . Note that yn converges to f 0 (p). Hence
1/n
f 0 (p) ∈ J. Therefore f 0 (I) ⊆ J. Since J ⊆ f 0 (I) ⊆ J and J being an interval, the
set f 0 (I) − J contains at most two elements, which are the end points of J. [5]

4. Let q, n be positive integers such that 1 < q < n and gcd(q, n) = 1.

a) Show that there exist unique integers k, r such that n = kq − r, 0 ≤ r < q.
b) Show that there exists a unique positive integer m and unique integers b1 , b2 , · · · , bm
n 1
all ≥ 2 satisfying = b1 −
q b2 − b31−
... 1
− 1 .
bm−1 − b
m
m
X
c) If bj > 2 for some j, then show that (bi − 2) < 2(n − q − 1). [13]
i=1
Solution:

(a) By division algorithm, there exists unique k1 , r1 ∈ N such that n = k1 q + r1 ,

where 0 ≤ r1 < q. Further, gcd(q, n) = 1 =⇒ r1 > 0.
Thus, we get, n = (k1 + 1)q − (q − r1 ). Since, r1 < q, we have, r = q − r1 >
0, r < q and let k = k1 + 1 ∈ N.
Hence, there exists unique k, r ∈ N such that n = kq − r, where 0 < r < q.
[3]
(b) By (a), we get k, r ∈ N such that n = kq − r and 0 < r < q.
n r
Thus, = k − . Let b1 = k. Now, n > q =⇒ k ≥ 2 =⇒ b1 ≥ 2.
q q

6
 n 1
Observe : = b1 − . Now, q > r; gcd(q, r) = 1. Let r1 = r. Hence, by
q q/r
applying the above procedure to q and r1 , we get,
q 1
= b2 −
r1 r1 /r2
where, b2 ≥ 2 and r2 < r1 .
Repeating these steps, we get, r1 , r2 , . . . , rm such that 0 < rm < rm−1 <
rm−2 < · · · < r2 < r1 . The process ends after finitely many steps(say m
steps), with rm = 1.
Now, we prove that, whenever, in the above expression, bi ≥ 2, for each
i = 1, 2, . . . , m then such bi s are unique. .j. . .k. . . . . (I)
n
In fact, we prove that if bi ≥ 2, for all i = 1, 2, . . . , m then b1 = +1
jnk q
Case (i) if b1 > + 1.
q
n r
In this case, we can write: = b1 − , for some r > q.
q q
n 1 1 q q
Thus, = b1 − = b1 − ; where b2 − s = , giving s = b2 −
q q/r b2 − s r r
Now, we have, r > q and b2 ≥ 2. Hence, s > 1.
n 1
Thus, = b1 − , where s > 1 . . . . . . (II)
q b2 − s
The above steps can be repeated to get s1 > 1 such that
n 1
= b1 −
q 1
b2 −
b 3 − s1
Hence, the process does not end after finitely many steps. This is a contra-
diction, since, wej have finitely many bi s.
nk
Case (ii) If b1 < + 1.
q
n r
In this case, we can write = b1 + , for some r > 0.
q q
n 1 1 q q
Thus, = b1 − = b1 − ; where b2 − s = − , giving s = b2 +
q −q/r b2 − s r r
Now, we have r > 0, q > 0 and b2 ≥ 2. Thus, s ≥ 2 > 1
n 1
Hence, we have, = b1 − with s > 1. Again as in case (i), the process
q b2 − s
does not end after finitely many j n k steps.
Hence, we must have b1 = + 1.
q
Now by replacing n by q and q by r, we get b2 is unique. Similarly we get
that all bi s are unique. [5]
(c) To prove that if bj > 2 for some j ∈ {1, 2, . . . , m} then
Xm
(bi − 2) < 2(n − q − 1).
i=1

7
We apply second principle of induction on n.
Observe that for n = 2, n = 3, n = 4, there is no value of q satisfying all
the conditions.
To prove for n = 5:
5 1
Case (i) q = 2. We have = 3 − and by (I), such expression is unique.
2 2
5 1
Case (ii) q = 3. We have = 2 − and by (I), such expression is unique.
3 3
Let n > 5. Assume the result is true for all k ≤ n.
To prove the result for n + 1.
Let 1 < q < n + 1 and gcd(n + 1, q) = 1
n+1 1 1
= b1 − = b1 −
q q/r b2 − b31−
.. 1
.− 1
bm−1 − b
m
q 1
By uniqueness of bi s, we have, = b2 −
r b3 −
... 1
− bm−1 − b1
. If no bi > 2 for
m
i = 2, 3, . . . , m then we get, bi = 2, for each i = 2, 3, . . . , m. So, we must
have b1 > 2.
n+1 1
In this case, we get = b1 − 1
q 2 − 2−
..
. − 2−1 1 .
2
1 k n+1 k−1
Observe that k−1
= . Hence = b1 − .
2− k k+1 q k
Now n + 1 = b1 q − (k−1) k
q. Thus k|q. Let q = km. Then we have
n + 1 = b1 q − (k − 1)m = b1 q − km + m = b1 q − q + m.
Therefore n − q = b1 q − 2q + m − 1 = q(b1 − 2) + (m − 1).
Thus we get 2(n + 1 − q − 1) = 2(n − q) = 2q(b1 − 2) + 2(m − 1) > b1 − 2
as required.

Now, if bi > 2, for some i = 2, 3, 4 . . . , m then applying induction hypothesis

to q, we get,
Xm
(bi − 2) < 2(q − r − 1)
i=2
m
X
We need to prove that (bi − 2) < 2(n + 1 − q − 1).
i=1
It is enough to prove that 2(q − r − 1) + b1 − 2 ≤ 2(n + 1 − q − 1)
That is to prove that 2q − 2r − 2 + b1 − 2 ≤ 2(b1 q − r − q − 1),
which is equivalent to b1 − 2 ≤ 2q(b1 − 2) which holds. Hence, the result is
true for n + 1. Thus by second principal of induction, the result is true for
all n ≥ 5. [5]

8
9