Sie sind auf Seite 1von 6

IMO Selection Test

Romania, 2004
Disclaimer: most of the proof here (except those which have no author indicated) are not
mine, and were taken from http://www.mathlinks.ro. Personally, I prefer solutions gathered
in one place, which is why I decided to write this le. The authors names are mathlinks
usernames. I do not have time now to check for their corectness, so please let me know if
something is wrong.
First Test
1. A quadrilateral of semiperimeter s has sides a
1
, a
2
, a
3
, a
4
. Prove that
4

1
1
s +a
i

2
9

1i<j4
1

(s a
i
)(s a
j
)
.
Find the case of equality.
Proof: (Arne) We have
2
9

1i<j4
1

(s a
i
)(s a
j
)

2
9
2

1i<j4
1
(s a
i
) + (s a
j
)
=
4
9

1i<j4
1
a
i
+a
j
.
Lets denote a
1
= a, a
2
= b, a
3
= c and a
4
= d. We should prove now that
2
9

1
a +b
+
1
a +c
+
1
a +d
+
1
b +c
+
1
b +d
+
1
c +d

1
3a +b +c +d
+
1
a + 3b +c +d
+
1
a +b + 3c +d
+
1
a +b +c + 3d
.
By Cauchy we have:

1
a +b
+
1
a +c
+
1
a +d

((a +b) + (a +c) + (a +d)) 9

1
9

1
a +b
+
1
a +c
+
1
a +d

1
3a +b +c +d
.
Adding the three analogous inequalities to this one we get the result.
2. A nite number of disjoint rectangles of total area 4 are placed in the plane so that
their projection on the horizontal axis is an interval. Prove that one can nd three
points in the union of the rectangles which are the vertices of a triangle of area at least
1.
1
3. Find all injective functions f : N N such that for all positive integers n we have
f(f(n))
n+f(n)
2
.
Proof: (Pbornsztein) Let f be a solution. Note that for each n we then have f(f(n))
maxn, f(n). Let f
k
denotes the k
th
iterate of f.
Suppose that there exists a such that a > f(a). Then f
2
(a) < a and by an easy
induction f
k
(a) < a for all k > 0 (1). Since there is a nite number of integers in
{1, 2, ..., a1} it follows that there exist 0 < i < j such that f
i
(a) = f
j
(a) = f
i
(f
ji
(a))
and using injectivity of f it leads to f
ji
(a) = a which contradicts (1).
Thus, a f(a) for each a > 0. Then f(a) f(f(a)) for each a. In another hand,
from the initial remark, we also have f(f(a)) f(a). Thus f(a) = f(f(a)) for each a,
and using injectivity of f we deduce that f(a) = a.
Conversely, it is easy to verify that f(x) = x is a solution.
4. Let D be a closed disc in the complex plane. Prove that for all positive integers n and
for all complex numbers z
1
, z
2
, . . . , z
n
D there exists a z D so that z
n
= z
1
z
2
z
n
.
Proof: (Andrew) We may use the following statement:
If f(z) is a polynom with roots u
1
, , u
n
, such that u
k
/ D for all k = 1, n and x is a
complex number in D, then all roots of g(z) = (x z) f

(z) + n f(z) are not in


D.
The proof is the following:
Let u is a root of g(z) and u in D, then (x u)f

(u) + nf(u) = 0
n
ux
=
f

(u)
f(u)
=
1
uu
1
+ +
1
uun
.
If t is inversion with center u, then t(x) is centre of mass of points t(u
1
), , t(u
n
),but
(D

= R
2
/t(D)-disk) t(x) / D

and t(u
1
), , t(u
n
) D

and D is a convex set.


We use lemma for f(z) = z
n
(z
1
)...(z
n
), x = z
1
, then: g(z) = (z
1
z) n z
n1
+
n(z
n
(z
1
) (z
n
)) = nz
1
(z
n1
(z
2
)...(z
n
)) all roots z
n1
(z
2
)...(z
n
) are not in D
and now we can use induction on n.
Second Test
1. Divide a closed disk in 2n equal ray sectors. Color n of them in black and n of them
in white. Number the white ones from 1 to n in clockwise direction and the black ones
the same way but in counterclockwise direction. Prove that there is a half disk made
of n such sectors, so that the half disk contains all numbers from 1 to n.
Proof: (Gorbber)
Hint:
We regard the sectors as colored and numbered points. The problem asks us to show
there is a line which contains points numbered on each of its sides.
2
We draw segments, each having points numbered the same as its extremities. Its
obvious that no such segments intersect. I think its easy to derive from here that
there must be a line cutting all segments, and that solves the problem.
(The owner of this site does not guarantee it works.)
2. Find all the positive integer values of the fraction
a
2
+ab +b
2
ab 1
,
where a, b are positive integers, not both equal to 1.
Proof: (Gorbber)
Hint:
We consider ordered pairs (a, b) s.t. a b. We can go from the pair (a, b) to a smaller
pair (a

, b

) = (
a+b+a
3
ab1
, a) until we eventually reach a dead-end, i.e. a pair which can
no longer become smaller by the above transformation.
A simple calculation shows that f(a, b) is invariant wrt the transformation mentioned,
so the value were interested in remains the same. I also think that the dead-end,
as Ive called it, must be one of the pairs (1, 2), (1, 4) and (2, 2). By analyzing the 3
cases we obtain f(a, b) {7, 4} (its 7 in the rst two cases and 4 in the third case).
(According to Valentin Vornicu who helps with the team selection, this does work.
Owner of sites hint: refer to IMO 1988 B3, IMO 1994 B1)
3. Consider three integers a, b, c with b odd. Let x
0
= 4, x
1
= 0, x
2
= 2c, x
3
= 3b and
x
n+3
= ax
n1
+bx
n
+cx
n+1
. Prove that p|x
p
m for all primes p and positive integers m.
Proof: (Alinbostan) Remark that x
n
equals the trace of the nth power of the (com-
panion) matrix
C =

0 0 0 a
1 0 0 b
0 1 0 c
0 0 1 0

,
and to apply (m times) Fermats little theorem
trace(C
p
) trace(C) (mod p).
(Author of solutions comment: American Mathematical Monthly, 105 April 1998,
problem 10655.)
4. The square ABCD lies in the interior of the circle C. Dene C
A
to be the circle that is
tangent to C, AB, AD and does not intersect with ABCD. Let A
1
be the intersection
of C and C
A
. Dene B
1
, C
1
, D
1
similarly. Prove that AA
1
, BB
1
, CC
1
, DD
1
intersect.
Proof: (Grobber) Let O
a
and O
b
be the centers of C
a
and C
b
respectively. Assume
these two circles touch AB at A
1
and B
1
. Let O be the center of the large circle and let
3
P be the center of ABCD. Assume OA

and OB

cut AB at A
2
and B
2
respectively.
Let X be A

AB.
Its easy to see that X is the external homothety center of C
a
and C
b
, so it also lies on
O
a
O
b
. This means that the lines A

, O
a
O
b
, A
2
B
2
and OX are concurrent in X, so
(A

, A
2
; O
a
, O) = (B

, B
2
; O
b
, O) (). Consider the homographic transformation from
the pencil of lines through A to the pencil of lines through B which maps AA

BB

,
AO
a
BO
b
and AO BO. From () we nd that it also maps AB = AA
2
BB
2
=
AB, so it has a xed element, and its well-known result that such transformations are
perspectivities (at least thats what theyre called in Romanian :)), i.e. the locus of the
intersection points of corresponding lines is a line, so O = AOBO, P = AO
a
BO
b
,
AA

BB

are collinear, and thats exactly what we wanted.


(Omid Hatami) You see M and N and P and Q are the points on the circle that
tangents from them to circle are respectively parallel to AB,BC,CD,DA. The tangents
from Q and M to intersect at A. The tangents from M and N to intersect at B. The
tangents from N and P to intersect at C. The tangents from P and Q to intersect
at D. Homothety from A takes A to A.So AA goes from A. So we must prove
AA,BB,CC and DD are concurrent. But this is obvious because sides of ABCD
and ABCD are parallel. And so theses lines are concurrent on a line connecting
center of square and circle. (Center of ABCD is center of circle).
Third Test
1. Let S be a set with at least 2 elements. Consider A
1
, A
2
, . . . , A
101
S so that the
union of any 50 such subsets has at least
50
51
|S| elements. Prove that there are 3 such
subsets, any two of which are nondisjoint.
Proof: (Harazi) Take a graph G in which you connect two non-disjoint subsets. Sup-
pose G doesnt have triangles. We prove it has at least 51 vertices of degree t most 50.
Otherwise, we can nd 51 vertices of degree at least 51. Take A one of them. Then
A must be connecte to a vertex of degree at least 51. Let it be B. Then its trivial to
show that we can nd C such that C is joined to A and B, contradiction. Thus, we
can nd A
1
, . . . , A
5
0 vertices of degree at most 50. Then A
1
is disjoint from at least
50 subsets, so it has at most
n
51
elements. The same for A
2
, . . . , A
5
0 and its done.
2. Prove that for all positive integers n, m so that m is odd we have
3
m
n|
m

k=0

3m
3k

(3n 1)
k
.
3. Let I be the incenter of the nonisosceles triangle ABC and let A

, B

, C

be the projec-
tions of I to BC, CA, AB. Let P be the intersection of the lines AA

, BB

, CC

. B

meets BC in N, A

meets AC in M. Prove that IP MN.


4
Proof: (Grobber) N is on BC and B

, which are the polars of A

and A respectively
(wrt the incircle). This means that AA

is the polar of N. In the same way we nd that


BB

is the polar of M, so P = AA

BB

must be the pole of MN, so PI MN (its


well-known that the polar of a point wrt a circle is perpendicular to the line joining
the point to the center of the circle).
4. Let n 2 be an integer and let a
1
, a
2
, . . . , a
n
be real numbers. Prove that for any
S {1, 2, 3, . . . , n} we have

iS
a
i

1ijn
(a
i
+ +a
j
)
2
.
Proof: (Andrei Negut) First, we will prove a lemma. For any real numbers a
1
, a
2
, ..., a
2k+1
we have
(a
1
+a
3
+a
5
+... +a
2k+1
)
2

1ijn
(a
i
+... +a
j
)
2
This is simple to prove, just by denoting a
1
+ a
2
+ ... + a
i
with S
i
and opening the
parenthesis. Now for our problem. Call a succession of consecutive a
i
s a sequence.
Call a sequence that is missing from S a gap. Thus, our set S will be formed by
a sequence, followed by a gap, then another sequence, then a gap,..., and at last a
sequence. That is why S will look like this:
{a
i
1
, a
i
1
+1
, ..., a
i
1
+
1
; a
i
2
, a
i
2
+1
, ..., a
i
2
+
2
; ...; a
i
k
, a
i
k
+1
, ..., a
i
k
+
k
}
where the
j
s are non-negative, and for every j we have i
j
+
j
< i
j+1
1. This is how
we write S as a sequence, then a gap, then another sequence,..., and at last a sequence.
Now let s
1
= a
i
1
+ a
i
1
+1
+ ... + a
i
1
+
1
, the sum of the rst sequence; s
2
= a
i
1
+
1
+1
+
... +a
i
2
1
, the sum of the rst gap; s
3
= a
i
2
+a
i
2
+1
+... +a
i
2
+
2
, the sum of the second
sequence; ... s
2k1
= a
i
k
+a
i
k
+1
+... +a
i
k
+
k
, the sum of the k-th sequence;
But then the sum in LHS becomes (s
1
+s
3
+... +s
2k1
)
2
, which the lemma tells us is
at most

1ij2k1
(s
i
+... +s
j
)
2
. The terms of this latter sum are obviously among
the terms in RHS, and thus follows that LHS RHS.
Fourth Test
1. Let m 2, n be integers. If n|a
m
1 for all a coprime to n, then prove that n
4m(2
m
1).
Proof: Let n = 2
k
r where r is odd. Modulo n we have that X
m
1 has at least (n)
distinct roots (where (n) is Eulers number) so m (n).
5
2. Let O be a point in the plane of the triangle ABC. A circle C which passes through
O intersects the lines OA,OB,OC in P, Q,R, respectively. The circle C also meets the
circumcircles of the triangles BOC, COA and AOB at K,L,M, respectively. Prove
that the lines PK,QL and RM intersect.
3. A convex polyhedron has n faces some of which are colored black so that no two black
faces have any points in common. The other faces are colored white. Prove that the
number of edges that belong to two white faces is at least n 2.
Fifth Test
1. Three circles K
1
, K
2
, K
3
of radii R
1
, R
2
, R
3
pass through the point O and intersect at
points A, B, C (K
2
K
3
= A, etc). The point O lies in the interior of the triangle
ABC. The line AO meets BC at A
1
and B
1
, C
1
are dened similarly.
If R is the circumradius of ABC then
R
OA
1
AA
1
R
1
+
OB
1
BB
1
R
2
+
OC
1
CC
1
R
3
.
2. An operation on an mn chess table consists in the following:
(a) Chose several unmarked squares, no two on a same row or column.
(b) Mark these squares with a 1.
(c) For any unmarked square which has a 1 on the same row or column, mark the
unmarked square with a 0.
A game consists of a number of operations. The game ends when no more operations
are possible. What is the maximum sum of the numbers on the table at the end of a
game.
3. Let p be a prime number and f Z[X] a polynomial
f(X) =
p1

i=1
a
i
X
i1
,
where a
i
= 1 if i is a square (mod p) and 1 otherwise.
(a) Prove that f(X) X 1 (mod (X 1)
2
), if p 3 (mod 4).
(b) Prove that f(X) (X 1)
2
(mod (X 1)
2
) if p 5 (mod 8)
6

Das könnte Ihnen auch gefallen