SEEMOUS 2008

South Eastern European Mathematical Olympiad

for University Students

Athens – March 7, 2008

Problem 1
Let f : [1, ∞) → (0, ∞) be a continuous function. Assume that for every a > 0, the
equation f (x) = ax has at least one solution in the interval [1, ∞).
(a) Prove that for every a > 0, the equation f (x) = ax has infinitely many solutions.
(b) Give an example of a strictly increasing continuous function f with these prop-
erties.

Problem 2
Let P0 , P1 , P2 , . . . be a sequence of convex polygons such that, for each k ≥ 0, the vertices
of Pk+1 are the midpoints of all sides of Pk . Prove that there exists a unique point lying
inside all these polygons.

Problem 3
Let Mn (R) denote the set of all real n × n matrices. Find all surjective functions f :
Mn (R) → {0, 1, . . . , n} which satisfy

f (XY ) ≤ min{f (X), f (Y )}

for all X, Y ∈ Mn (R).

Problem 4
Let n be a positive integer and f : [0, 1] → R be a continuous function such that
Z 1
xk f (x) dx = 1
0

for every k ∈ {0, 1, . . . , n − 1}. Prove that

Z 1
(f (x))2 dx ≥ n2 .
0

1
 Answers

Problem 1
Solution. (a) Suppose that one can find constants a > 0 and b > 0 such that f (x) 6= ax
for all x ∈ [b, ∞). Since f is continuous we obtain two possible cases:
1.) f (x) > ax for x ∈ [b, ∞). Define

f (x) f (x0 )
c = min = .
x∈[1,b] x x0

Then, for every x ∈ [1, ∞) one should have

min(a, c)
f (x) > x,
2
a contradiction.
2.) f (x) < ax for x ∈ [b, ∞). Define

f (x) f (x0 )
C = max = .
x∈[1,b] x x0

Then,
f (x) < 2 max(a, C)x
for every x ∈ [1, ∞) and this is again a contradiction.

(b) Choose a sequence 1 = x1 < x2 < · · · < xk < · · · such that the sequence
yk = 2k cos kπ xk is also increasing. Next define f (xk ) = yk and extend f linearly on
each interval [xk−1 , xk ]: f (x) = ak x + bk for suitable ak , bk . In this way we obtain an
increasing continuous function f , for which lim f (x 2n )
x2n = ∞ and lim
f (x2n−1 )
x2n−1 = 0. It
n→∞ n→∞
f (x)
now follows that the continuous function x takes every positive value on [1, ∞).

Problem 2
Solution. For each k ≥ 0 we denote by Aki = (xki , yik ), i = 1, . . . , n the vertices of Pk .
We may assume that the center of gravity of P0 is O = (0, 0); in other words,
1 0 1
(x1 + · · · + x0n ) = 0 and (y10 + · · · + yn0 ) = 0.
n n

Since 2xk+1
i = xki + xki+1 and 2yik+1 = yik + yi+1
k for all k and i (we agree that xkn+j = xkj
k
and yn+j = yjk ) we see that

1 k 1
(x1 + · · · + xkn ) = 0 and (y1k + · · · + ynk ) = 0
n n
for all k ≥ 0. This shows that O = (0, 0) is the center of gravity of all polygons Pk .
In order to prove that O is the unique common point of all Pk ’s it is enough to prove
the following claim:
Claim. Let Rk be the radius of the smallest ball which is centered at O and contains Pk .
Then, lim Rk = 0.
k→∞

2
Proof of the Claim. Write k · k2 for the Euclidean distance to the origin O. One can
easily check that there exist β1 , . . . , βn > 0 and β1 + · · · + βn = 1 such that
n
X
Ak+n
j = βi Akj+i−1
i=1
Pn k
for all k and j. Let λ = min βi . Since O = i=1 Aj+i−1 , we have the following:
i=1,...,n
° °
°Xn °
° °
kAk+n
j k2 = ° (βi − λ)Akj+i−1 °
° °
i=1 2
n
X
≤ (βi − λ)kAkj+i−1 k2
i=1
n
X
≤ Rk (βi − λ) = Rk (1 − nλ).
i=1

This means that Pk+n lies in the ball of radius Rk (1 − nλ) centered at O. Observe that
1 − nλ < 1.
Continuing in the same way we see that Pmn lies in the ball of radius R0 (1 − nλ)m
centered at O. Therefore, Rmn → 0. Since {Rn } is decreasing, the proof is complete.

Problem 3
Solution. We will show that the only such function is f (X) = rank(X). Setting
Y = In we find that f (X) ≤ f (In ) for all X ∈ Mn (R). Setting Y = X −1 we find
that f (In ) ≤ f (X) for all invertible X ∈ Mn (R). From these facts we conclude that
f (X) = f (In ) for all X ∈ GLn (R).
For X ∈ GLn (R) and Y ∈ Mn (R) we have
f (Y ) = f (X −1 XY ) ≤ f (XY ) ≤ f (Y ),
f (Y ) = f (Y XX −1 ) ≤ f (Y X) ≤ f (Y ).
Hence we have f (XY ) = f (Y X) = f (Y ) for all X ∈ GLn (R) and Y ∈ Mn (R). For
k = 0, 1, . . . , n, let µ ¶
Ik O
Jk = .
O O
It is well known that every matrix Y ∈ Mn (R) is equivalent to Jk for k = rank(Y ).
This means that there exist matrices X, Z ∈ GLn (R) such that Y = XJk Z. From
the discussion above it follows that f (Y ) = f (Jk ). Thus it suffices to determine the
values of the function f on the matrices J0 , J1 , . . . , Jn . Since Jk = Jk · Jk+1 we have
f (Jk ) ≤ f (Jk+1 ) for 0 ≤ k ≤ n − 1. Surjectivity of f imples that f (Jk ) = k for
k = 0, 1, . . . , n and hence f (Y ) = rank(Y ) for all Y ∈ Mn (R).

Problem 4
Solution. There exists a polynomial p(x) = a1 + a2 x + · · · + an xn−1 which satisfies
Z 1
(1) xk p(x) dx = 1 for all k = 0, 1, . . . , n − 1.
0

3
It follows that, for all k = 0, 1, . . . , n − 1,
Z 1
xk (f (x) − p(x)) dx = 0,
0

and hence Z 1
p(x)(f (x) − p(x)) dx = 0.
0
Then, we can write
Z 1 Z 1
2
(f (x) − p(x)) dx = f (x)(f (x) − p(x)) dx
0 0
Z 1 n−1
X Z 1
2
= f (x) dx − ak+1 xk f (x) dx,
0 k=0 0

and since the first integral is non-negative we get

Z 1
f 2 (x) dx ≥ a1 + a2 + · · · + an .
0

To complete the proof we show the following:

Claim. For the coefficients a1 , . . . , an of p we have

a1 + a2 + · · · + an = n2 .

Proof of the Claim. The defining property of p can be written in the form
a1 a2 an
+ + ··· + = 1, 0 ≤ k ≤ n − 1.
k+1 k+2 k+n
Equivalently, the function
a1 a2 an
r(x) = + + ··· + −1
x+1 x+2 x+n
has 0, 1, . . . , n − 1 as zeros. We write r in the form

q(x) − (x + 1)(x + 2) · · · (x + n)
r(x) = ,
(x + 1)(x + 2) · · · (x + n)

where q is a polynomial of degree n − 1. Observe that the coefficient of xn−1 in q is

equal to a1 + a2 + · · · + an . Also, the numerator has 0, 1, . . . , n − 1 as zeros, and since
lim r(x) = −1 we must have
x→∞

q(x) = (x + 1)(x + 2) · · · (x + n) − x(x − 1) · · · (x − (n − 1)).

n(n+1)
This expression for q shows that the coefficient of xn−1 in q is 2 + (n−1)n
2 . It follows
that
a1 + a2 + · · · + an = n2 .