INMO TEST # 1_Solutions
INMO TEST # 1_Solutions
INMO
TARGET INMO
Instructions : Time : 3 Hours
8 Calculators (in any form) and protractors are not allowed.
8 Rulers and compasses are allowed.
8
8 Pre Nurture and Career Foundation Division
Answer all the questions.
All questions carry equal marks.
8 Answer to each question should start on a new page. Clearly indicate the question number.
1. A subset of {1, 2, 3, 4,----, n} is said to be amazing if all the elements in that subset are greater-
than or equal to the cardinality of the subset. If Tn is the number of such amazing subsets, then
n
prove that åT
i =1
i
2
is divisible by Tn+1.
2. Given a positive integer k, show that there exists a prime p such that one can choose distinct
integers a1, a2, ….ak+3 Î {1, 2, …., p – 1} such that p divides aiai+1ai+2ai+3 – i for all i = 1, 2, 3, …,
k.
Sol. First we choose distinct positive rational numbers r1,…,rk+3 such that
riri+1 ri+2 ri+3 = i for 1 £ i £ k
Let r1 = x, r2 = y, r3 = z be some distinct primes greater than k; the remaining terms satisfy
1 i +1
r4 = and ri + 4 = ri . It follows that if ri are represented as irreducible fractions, the
r1 r2 r3 i
numerators are divisible by x for i º 1 (mod 4), by y for i º 2 (mod 4), by z for i º 3 (mod 4) and
by none for i º 0 (mod 4). Notice that ri < ri +4 thus the sequences
r1 < r5 < r9 < ……, r2 < r6 < r10 < ….., r3 < r7 < r11 < ….., r4 < r8 < r12 < ……
are increasing and have no common terms, that is, all ri are distinct.
u
If each ri is represented by an irreducible fraction i , choose a prime p which divides neither
vi
vi, 1 £ i £ k + 1, nor vivj (ri – rj) = vjui – viuj for i < j, and define ai by the congruence a i vi º u i
(mod p). Since ri ri+1 ri+2 rr+3 = i, we have
ivi vi +1vi + 2 vi +3 º ri vi ri +1vi +1ri + 2 vi + 2 ri +3v i+ 3
u i u i +1u i+ 2 u i +3 º a i vi a i vi +1a i + 2 vi + 2a i +3 vi +3 (mod p)
ant therefore ai ai+1ai+2ai+3 º i(modp) for 1 £ i £ k.
If ai º aj (mod p), then uivj º aivivj º ujvi (mod p), a contradiction.
3. Let M, N, K be points chosen on the sides AB, BC, AC of triangle ABC, respectively. Given that
triangle MNK is an acute triangle and that hm + hn + hk – min(hm,hn,hk) = min(ha,hb,hc) where ha,
hb, hc and hm, hn, hk are the altitudes of triangles ABC and MNK, respectively. Find all possible
N, K are the midpoints of the sides of triangle ABC. Hence, if follows that
( ABC) = 4
( MNK )
4. A game is played with 16 cards laid out in a row. Each card has a black side and a red side, and
initially the face-up sides of the cards alternate black and red with the leftmost card black-side-
up. A move consists of taking a consecutive sequence of cards (possibly only containing 1 card)
with leftmost card black-side-up and the rest of the cards red-side-up, and flipping all of these
cards over. The game ends when a move can no longer be made. What is the maximum possible
number of moves that can be made before the game ends?
Sol. 215 + 213 + 211 + 29 + 27 + 25 + 23 + 21 = 43690
For each arrangement of cards, consider the corresponding binary number constructed by
assigning a 1 to each black card and a 0 to each red card. Then the operation decreases the binary
number strictly, so the answer is at most 10101010101010102 = 43690. Note that this is
achievable if we go down by 1 each time by picking the rightmost black card, so this is the
answer.
5. Let ABCD be a quadrilateral, such that ÐA = 75°, ÐB = 45°, ÐC = 120°. Given that lines AB,
CD intersect at point E and lines AD, BC intersect at point F. Denote by M, N, P, Q the
intersection points of the altitudes of triangles EAD, EBC, FCD, FAB respectively. Find
2
æ MN ö
ç PQ - 3 ÷ .
è ø
Sol. Let CC1, DD1 be the altitudes of triangle FCD. Note that
FD1 FC
=cos ÐAFB = 1
FD FC
Thus, it follows that
DC1FD1 ~ DCFD …… (1)
Therefore
C1D1 = CD cos ÐAFB
On the other hand points F, D1, P, C1 are on the circle with diameter FP. Hence, according to the
law of sines, we obtain that
C1D1
FP = …… (2)
sin ÐAFB
From (1) and (2), we deduce that
FP = CD cot ÐAFB, Similarly FQ = AB cot ÐAFB
Now, let us consider a parallelogram ADCK. Note that FP ^ AK, FQ ^ AB. Therefore ÐPFQ =
ÐKAB. On the other hand,
FP FQ
=
AK AB
Thus, it follows that
DPFQ ~ DKAB, Hence PQ = KB cot ÐAFB
In a similar way, we obtain that
MN = KB cot ÐBEC
We deduce that
MN cot15°
=
PQ cot 60°
= 3 2+ 3 ( )
2
æ MN ö
Therefore ç PQ - 3 ÷ = 12
è ø
6. Let a be a non-zero real number. Find all functions ƒ : ¡® ¡ such that
xƒ(x + y) = (x + ay)ƒ(x) + xƒ(y)
for all x, y Î ¡ .
Sol. ƒ(x) = cx2 for any real constant c if a = 2; ƒ(x) = 0 otherwise.
Let P(x,y) denote the assertion of the given functional equation. Note that P(1, 0) is
ƒ(1) = ƒ(1) + ƒ(0) which implies
ƒ(0) = 0
Applying this result to P(x, –x) and P(–x, x) where x ¹ 0 we get :
0 = (1 – a)x ƒ(x) + xƒ(–x) (1)
0 = (a – 1)x ƒ(–x) – x ƒ(x) (2)
By adding (1) and (2) and simplifying, we get 0 = axƒ(–x) – axƒ(x) which implies
ƒ(x) = ƒ(–x)
for all x ¹ 0. Since ƒ(0) = 0 = ƒ(–0), we can conclude that ƒ is even. Therefore (1) simplifies to
0 = x ƒ(x) (2 – a)
which implies that if a ¹ 2 then ƒ(x) = 0 for all x Î ¡ . It is easy to check that this function
works.
Now let us consider the case a = 2. The initial functional equation becomes
xƒ(x + y) = (x + 2y)ƒ(x) + xƒ(y)
which can be rewritten as
xƒ(x + y) – (x + y) ƒ(x) =yƒ(x) + xƒ(y).
Note that the right hand side is symmetric with respect to x and y. From this we can duduce that
xƒ(x + y) – (x + y) ƒ(x) = yƒ(x + y) – (x + y) ƒ(y) where factorizing yields
(x – y) ƒ (x + y) = (x + y) (ƒ(x) – ƒ(y)) (3)
By replacing y with –y and using the fact that ƒ is even, we get
(x + y) ƒ (x – y) = (x – y) (ƒ(x) – ƒ(y)) (4)
z +1 z -1
Taking x = and x = in both (3) and (4), we get
2 2
æ æ z +1 ö æ z -1 ö ö
ƒ ( z ) =z ç ƒ ç ÷ -ƒç ÷÷ (5)
è è 2 ø è 2 øø
æ z +1 ö æ z -1 ö
zƒ (1) =ƒ ç ÷-ƒç ÷ (6)
è 2 ø è 2 ø
respectively. Equations (5) and (6) together yield ƒ(z) = z.zƒ(1) = z2ƒ(1) which must hold for all
zΡ .
Thus, the only possible functions that satisfy the given relation for a = 2 are ƒ(x) = cx2 for some
real constant c. it is easy to check that they indeed work.
Sol2. Multiplying the given equation by y gives
xy ƒ(x + y) = (x + ay) y ƒ (x) + xy ƒ(y)
which is equivalent to
xy (ƒ(x + y) – ƒ(x) – ƒ(y)) = ay2 ƒ(x)
The left-hand side of this equation is symmetric in x and y. Hence the right-hand side must also
stay the same if we swap x and y, i.e.,
ay2 ƒ(x) = ax2 ƒ(y)
As a ¹ 0, this implies
y2 ƒ(x) = x2 ƒ(y)
Setting y = 1 in this equation immediately gives ƒ(x) = cx2 where c = ƒ(1). Plugging ƒ(x) = cx2
into the original equation gives
cx (x + y)2 = c(x + ay)x2 + cxy2
where terms can be rearranged to obtain
cx(x + y)2 = cx(x2 + axy + y2)
If c = 0 then (7) is satisfied. Hence for every a, the function ƒ(x) = 0 is a solution. If c ¹ 0 then
(7) is satisfied if and only if a = 2. Hence in the case a = 2, all functions ƒ(x) = cx2 (where c ¹ 0)
are also solutions.