40th International Mathematical Olympiad

Bucharest, Romania, July 1999.

1. Determine all finite sets S of at least three points in the plane which satisfy the
following condition:
for any two distinct points A and B in S, the perpendicular bisector of the line
segment AB is an axis of symmetry for S.

Soln. Let G be the centre of gravity of the set S. Since the perpendicular bisector of
the line joining every pair of points is an axis of symmetry, G lies on the perpendicular
bisector. Thus the perpendicular bisectors meet at a common point G. Thus the points lie
on a circle. Let a, b, c be three consecutive points. Since the perpendicular of ac is an axis
of symmetry, b must lie on it. Thus the lengths of ab and bc are equal. Thus the points
are the vertices of a regular polygon.
Second soln. One can show easily that the boundary of the convex hull of the points
form a regular polygon as in the second half of the previous proof. The only thing left to
do is to prove that there is no point in the interior. This can be proved by contradiction.
2. Let n be a fixed integer, with n 2.
(a) Determine the least constant C such that the inequality
X

xi xj (x2i + x2j ) C

1i<jn

 X

xi

4

1in

holds for all real numbers x1 , . . . , xn 0.

(b) For this constant C, determine when equality holds.
Soln. The inequality
is symmetric and homogeneous, so we can assume x1 x2
P
xn 0 and
xi = 1. In this case we have to maximize the sum
F (x1 , . . . , xn ) =

xi xj (x2i + x2j ).

i<j

Let xk+1 be the last nonzero coordinate and we assume that k 2. We shall replace
x = (x1 , . . . , xk , xk+1 , 0, . . . , 0) with
x0 = (x1 , . . . , xk1 , xk + xk+1 , 0, . . . , 0)
1

to increase the value of F as shown below:

0

F (x )F (x) = xk xk+1 [3(xk + xk+1 )

k1
X

xi x2k x2k+1 ]

i=1

= xk xk+1 [3(xk + xk+1 )(1 xk xk+1 ) x2k x2k+1 ]

= xk xk+1 (xk + xk+1 )(3 4(xk + xk+1 )) + 2xk xk+1 ].
From

1
(xk + xk+1 ) + xk + xk+1
2
it follows that 2/3 xk + xk+1 , and therefore F (x0 ) F (x) > 0. After several such
substitutions, we have
1 x1 + xk + xk+1

F (x) F (a, b, 0, . . . , 0) = ab(a2 + b2 )

1
1
= (2ab)(1 2ab) = F (1/2, 1/2, 0, . . . , 0).
2
8
Thus C = 1/8 and equality occurs if and only if two of the xi s are equal (possibly zero)
and the remaining variables are zero.
3. Consider an n n square board, where n is a fixed even positive integer. The board is
divided into n2 unit squares. We say that two different squares on the board are adjacent
if they have a common side. N unit squares on the board are marked in such a way that
every square (marked or unmarked) on the board is adjacent to at least one marked square.
Determine the smallest possible value of N .
Soln. Let n = 2k. First colour the board black and white like a chessboard. We say that
a set S of cells is covered by a set M of marked cells if every cell in S is neighbour to a
cell in M . Let fw (n) be the minimal number of white cells that must be marked so as to
cover all the black cells. Define similarly fb (n). Due to symmetry of the chessboard (n is
even), we have
fw (n) = fb (n).
Since black cells can only be covered by marked white cells and vice versa, we have
N = fw (n) + fb (n).
Place the board so that the longest black diagonal is horizontal. Now consider the horizontal rows of white cells. Mark the odd cells of the first row, third row, fifth row, etc.
(The figure illustrates the case n = 8). Call this set of cells M . The number of cells in M
is
k(k + 1)
1 + 2 + + k =
.
2
It is easy to see that the black cells are covered by M . Thus
fw (n)

k(k + 1)
2
2

.....
..... .....
.... .....
......................................................
.
.
.
.
.
.
...........................................
..... ......... ......... ....
........ .......... .......... .........
...........................................................................................................................
.
.
.
.
.
. . ...... . ...... . .... ........ .
..... .................. .................. .................. ................. .........
..
..
......
..
...
...
............ ........... ............ ............ ........... ...........
..............................................................................................................................................................................................................................
.
.
.
. ............ .......... ......... .......... .......... ........... ....
.
.
.
.
.
.
.
.
............... .............. ............... ............... ............... .............. ............... ................
...................................................................................................................................................................................................................................................................
............ ............ ............ ............ ............ ............ .......... ............
...... ....... ........ ........ ........ ........ ........ .....
..... ................ ................ ................ ................ ................ ................ .....
.............................................................................................................................
............... ................. ............... ............... .............. .................
..
..
...
.......
..
..
..... .................. ..................... ..................... .................... ........
........................................................................................
.................. ................. ................. ..................
......... ........ ....... .........
..... ........... ............ .....
..............................................
........................................
........... ...........
..... .....
..........
..

x
Now consider the cells in M . No two of them have a common black neighbour. So we
need to mark at least k(k + 1)/2 black cells in order to cover M . Therefore
fb (n)

k(k + 1)
.
2

Hence
fb (n) = fw (n) = k(k + 1)/2.
Thus
N = fb (n) + fw (n) = k(k + 1).

4. Determine all pairs (p, q) of positive integers such that p is prime, n 2p, and
(p 1)n + 1 is divisible by np1 .
Soln. Clearly (1, p) and (2, 2) are solutions and for other solutions we have p 3. Now
assume that n 2 and p 3. Since (p 1)n + 1 is odd and is divisible by np1 , n must
be odd. Thus n < 2p. Let q be the smallest prime divisor of n. From q | (p 1)n + 1, we
have
(p 1)n 1 (mod q) and gcd(q, p 1) = 1.
But gcd(n, q 1) = 1 (from the choice of q), there exist integers u and v such that
un + v(q 1) = 1, whence
p 1 (p 1)un (p 1)v(q1) (1)u 1v 1 (mod q),
because u must be odd. This shows q | p and therefore q = p. Hence n = p. Now
pp1 | (p 1)p + 1

 





p p3
p
p
2
p2
=p p

p
+ +
p
+1
1
p3
p2
Since every term in the bracket except the last is divisible by p, we have p 1 2. Thus
p = 3 = n. Indeed (3, 3) is a solution.
In conclusion, the only solutions are (1, p), (2, 2), (3, 3).
5. Two circles 1 and 2 are contained inside the circle , and are tangent to at the
distinct points M and N , respectively. 1 passes through the centre of 2 . The line passing
3

through the two points of intersection of 1 and 2 meets at A and B. The lines M A
and M B meet 1 at C and D, respectively.
Prove that CD is tangent to 2 .
Soln. We first prove a lemma.
Lemma: Let 1 and 2 be two circles such that one does not contain the other. Let
be a circle containing both 1 , 2 and touches 1 at P , 2 at S. Let ` be a chord of
which is tangent to 1 at Q and 2 at R, with both circles on the same side of the chord.
Then P Q and SR meet at a point M which is on . Furthermore M is on the radical axis
of 1 and 2 .
........................................
.
. ..........
.........
.......
.......
.
......
.......................
..
.....
.
.
.
.
.
.
.....
.....
..
.
.
.
.
.
.....
...
.........
.
.
.
.
.
.
...
.
.... .... ....... 1 ....
.
...
.
.
. ... .. .....
..
...
.
.
.
.....
. ...
......... ... ...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
........ ....
..... ..
....... .. ..
...
.
.
.
.
.
.
.
.
.
.
.
..... ...
..... .. .
.
. ........... ...
.
.
.
.
.... ...
.
.
.
.
.
.
.. .
........
.
.....
...
...
........................................ ..
......
..
...
.....
... ........
.
....
.
.
.
....
.
.
......
...
. ...... ..
...
.
.
2
..
.
..... .
......
...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...
.
.
.......
........
...
.
.
...
.
.
.
.
.
...
.
.
......
.. ...
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...
...... ...
..
...
....
.
...... ...
.......
...
...
.....
........
.......
...
...
...
.. .
......... ..............
...
.........
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
...
.
... ..
.. .
....... ..................................... .....
...
...
.......
......
...
..
.......
......
... .... .....
...
.......
.
.
.
.
.
.
.
..... ...
.
.
.
.
.
..........
..... .... ............
.
.
... ....
................
..... ...........
......
......
.......
.....
.......
.........
.................................................

PO

M
Proof : Let O, O1 , O2 be the centres of , 1 , 2 , respectively. Produce P Q to meet
at X. Then it is easy to see that O1 Q||OX. Thus the tangent at X is parallel to `.
Produce SR to meet at Y . A similar consideration shows that the tangent at Y is
also parallel to `. Thus X = Y = M . Since ` is parallel to the tangent at M , we have
SP Q = 180 QRS. Thus P QRS is cyclic. Hence the powers of M with respect to
1 and 2 are equal and M is on the radical axis of both. This completes the proof of the
lemma.
Let E be the intersection of N A with 2 and F be the intersection of N B with 2 .
From the lemma, and since the radical axis in this case is the common chord AB, we
known that CE and DF are both common tangents of 1 and 2 . Thus O1 O2 is the
perpendicular bisector of CD, i.e., O2 C = O2 D and O2 CD = O2 DC. Since CE is
tangent to 1 , ECO2 = O2 DC = O2 CD. Thus CD is tangent to 2 .
A
.......... C .......
...
.............. ......... .............
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..............
. ..
.......
...................... .... ........
........
.
.
.......
.
.
.
.
.
.
.
.
.
.
.
.
.
.
... .................. ......E
...
.
.
...
.
.
.
.
.
.
.
.
.
.
.
.
... ... ..........................................
....
.. .................
...............................
M........................................
..... . .
..... ....
... ....... ......... .....
....
..... ..
.... ..
.. ...
... ........
........
.
.
.
....... ...
.....
.......
..
... ..
......
.....
.
O
.
....
... ......
1
...
.....
..
...
....... N
...
.....
... .
.
.
.
.
.

O
.
.
.
.
.
.....
...
... .....
....
2 .... ....
.....
...
... ... ...
......
.....
.. ..
...
... .
.
.....
... ...
... ..... ........ .....
...
.
.
... ......
.....
...
.. ...... ... ... ...
.
.
.
.
.
....... ....
.....
..
...
. .. ..
.................
.....
....
..... .... .... .................................................
.....
......
..... . ... .................. ..
... .. .......... ..
1 ...........
..
............... .. ............................. ....
...
.............. .....
... ....
.....
.
.
..........
.
..... ... ...
..... .. ...
.........
.......

6. Determine all functions f : R R such that

f (x f (y)) = f (f (y)) + xf (y) + f (x) 1
4

for all x, y R.
Soln. Let A = Imf and c = f (0). By putting x = y = 0, we get f (c) = f (c) + c 1, so
c 6= 0.
It is easy to find the restriction f to A. Take x = f (y) to obtain
f (x) =

c + 1 x2

2
2

for all x A.

(1)

The main step is to show that A A = R. Indeed, for y = 0, we get:

{f (x c) f (x) | x R} = {cx + f (c) 1 | x R} = R
because c 6= 0.
With this we conclude that the given functional equation is equivalent to the following:
f (x y) = f (y) + xy + f (x) 1,

for all x, y R.

By putting y = 0, we get f (0) = 1. Replacing y by x, we have

1 = 2f (x) + x2 1,

i.e., f (x) = 1

x2
.
2

It is easy to check that this function satisfies the given functional equation.