is a Cauchy sequence in H. (This is the crucial observation.

) Indeed, for m > n, the norm kym − yn k IMC2008, Blagoevgrad, Bulgaria

may be computed by the above remark as Day 2, July 28, 2008
⊤ 2
d2 d2 n(m − n)2 m − n
  
1 1 1 1 1 1
kym − yn k2 = − ,..., − , ,..., = +
2 m n m n m m 2 m2 n2 m2 Problem 1. Let n, k be positive integers and suppose that the polynomial x2k − xk + 1 divides
Rm
2
d (m − n)(m − n + n) 2
d m−n d 2

1 1
 x2n + xn + 1. Prove that x2k + xk + 1 divides x2n + xn + 1.
= = = − → 0, m, n → ∞. Solution. Let f (x) = x2n + xn + 1, g(x) = x2k − xk + 1, h(x) = x2k + xk + 1. The complex number
2 m2 n 2 mn 2 n m
π π
x1 = cos( 3k ) + i sin( 3k ) is a root of g(x).
By completeness of H, it follows that there exists a limit πn n
Let α = 3k . Since g(x) divides f (x), f (x1 ) = g(x1 ) = 0. So, 0 = x2n 1 + x1 + 1 = (cos(2α) +
y = lim yn ∈ H. i sin(2α)) + (cos α + i sin α) + 1 = 0, and (2 cos α + 1)(cos α + i sin α) = 0. Hence 2 cos α + 1 = 0, i.e.
n→∞
α = ± 2π + 2πc, where c ∈ Z.
We claim that y sastisfies all conditions of the problem. For m > n > p, with n, p fixed, we compute 3
3k −1
Let x2 be a root of the polynomial h(x). Since h(x) = xxk −1 , the roots of the polynomial h(x)
⊤ 2
d2 are distinct and they are x2 = cos 2πs + i sin 2πs

1 1 1 1 1 , where s = 3a ± 1, a ∈ Z. It is enough to prove that
kxn − ym k2 = − ,...,− ,1 − ,− ,...,− 3k
n
3k
2 m m m m m f (x2 ) = 0. We have f (x2 ) = x2n 2 + x2 + 1 = (cos(4sα) + sin(4sα)) + (cos(2sα) + sin(2sα)) + 1 =
Rm
2 2 2 (2 cos(2sα) + 1)(cos(2sα) + i sin(2sα)) = 0 (since 2 cos(2sα) + 1 = 2 cos(2s(± 2π + 2πc)) + 1 =
d2
 
d m − 1 (m − 1) d m−1 3
= + = → , m → ∞, 2 cos( 4πs
3
) + 1 = 2 cos( 4π 3
(3a ± 1)) + 1 = 0).
2 m2 m2 2 m 2
√ Problem 2. Two different ellipses are given. One focus of the first ellipse coincides with one focus
showing that kxn − yk = d/ 2, as well as of the second ellipse. Prove that the ellipses have at most two points in common.
* ⊤
d2 1 1 1 1 Solution. It is well known that an ellipse might be defined by a focus (a point) and a directrix (a
hxn − ym , xp − ym i = − ,...,− ,...,1 − ,...,− , straight line), as a locus of points such that the distance to the focus divided by the distance to
2 m m m m
 ⊤ + directrix is equal to a given number e < 1. So, if a point X belongs to both ellipses with the same
1 1 1 1 focus F and directrices l1 , l2 , then e1 · l1 X = F X = e2 · l2 X (here we denote by l1 X, l2 X distances
− ,...,1 − ,...,− ,...,−
m m m m between the corresponding line and the point X). The equation e1 · l1 X = e2 · l2 X defines two lines,
Rm
whose equations are linear combinations with coefficients e1 , ±e2 of the normalized equations of lines
d2 m − 2 d2
  
2 1
= − 1− =− → 0, m → ∞, l1 , l2 but of those two only one is relevant, since X and F should lie on the same side of each directrix.
2 m2 m m 2m
So, we have that all possible points lie on one line. The intersection of a line and an ellipse consists
showing that hxn − y, xp − yi = 0, so that of at most two points.
(√ )
2 Problem 3. Let n be a positive integer. Prove that 2n−1 divides
(xn − y) : n ∈ N
d X  n 
5k .
is indeed an orthonormal system of vectors. 2k + 1
0≤k<n/2
This completes the proof in the case when T = S, which we can always take if S is countable. If
it is not, let x′ , x′′ be any two distinct points in S \ T . Then applying the above procedure to the set  √ n  √ n 
Solution. As is known, the Fibonacci numbers Fn can be expressed as Fn = √15 1+ 5
2
− 1−2 5 .
′ ′ ′′
T = {x , x , x1 , x2 , . . . , xn , . . .} 
l−1

n n n
1



Expanding this expression, we obtain that Fn = 2n−1 1 + 3 5 + ... + l 5 2 , where l is the
it follows that
greatest odd number such that l ≤ n and s = l−1 ≤ n2 .
x′ + x′′ + x1 + x2 + · · · + xn x1 + x2 + · · · + xn s
2
lim = lim =y n

k n


1
5 , which implies that 2n−1 divides 0≤k<n/2 5k .
P P
n→∞ n+2 n→∞ n So, Fn = 2n−1 2k+1 2k+1
k=0
satisfies that (√ √ ) (√
Problem 4. Let Z[x] be the ring of polynomials with integer coefficients, and let f (x), g(x) ∈ Z[x] be
)
2 ′ 2 ′′ 2
(x − y), (x − y) ∪ (xn − y) : n ∈ N nonconstant polynomials such that g(x) divides f (x) in Z[x]. Prove that if the polynomial f (x)−2008
d d d
has at least 81 distinct integer roots, then the degree of g(x) is greater than 5.
is still an orthonormal system. Solution. Let f (x) = g(x)h(x) where h(x) is a polynomial with integer coefficients.
This it true for any distinct x′ , x′′ ∈ S \ T ; it follows that the entire system Let a1 , . . . , a81 be distinct integer roots of the polynomial f (x) − 2008. Then f (ai ) = g(ai )h(ai ) =
(√ )
2 2008 for i = 1, . . . , 81, Hence, g(a1 ), . . . , g(a81 ) are integer divisors of 2008.
(x − y) : x ∈ S Since 2008 = 23 ·251 (2, 251 are primes) then 2008 has exactly 16 distinct integer divisors (including
d
the negative divisors as well). By the pigeonhole principle, there are at least 6 equal numbers among
is an orthonormal system of vectors in H, as required. g(a1 ), . . . , g(a81 ) (because 81 > 16 · 5). For example, g(a1) = g(a2 ) = . . . = g(a6 ) = c. So g(x) − c is

4 1
a nonconstant polynomial which has at least 6 distinct roots (namely a1 , . . . , a6 ). Then the degree Problem 6. Let H be an infinite-dimensional real Hilbert space, let d > 0, and suppose that S is a
of the polynomial g(x) − c is at least 6. set of points (not necessarily countable) in H such that the distance between any two distinct points
Problem 5. Let n be a positive integer, and consider the matrix A = (aij )1≤i,j≤n , where in S is equal to d. Show that there is a point y ∈ H such that
(√ )
( 2
1 if i + j is a prime number, (x − y) : x ∈ S
aij = d
0 otherwise.
is an orthonormal system of vectors in H.
Prove that | det A| = k 2 for some integer k. √
Solution. It is clear that, if B is an orthonormal system in a Hilbert space H, then {(d/ 2)e : e ∈ B}
Solution. Call a square matrix of type (B), if it is of the form is a set of points in H, any two of which are at distance d apart. We need to show that every set S
  of equidistant points is a translate of such a set.
0 b12 0 ... b1,2k−2 0 We begin by noting that, if x1 , x2 , x3 , x4 ∈ S are four distinct points, then
 b21 0 b23 ... 0 b2,2k−1 
hx2 − x1 , x2 − x1 i = d2 ,
 
 0 b32 0 ... b3,2k−2 0 
 .. .. .. .. .. .
 
.. 1
1
 . . . . . .  hx2 − x1 , x3 − x1 i = kx2 − x1 k2 + kx3 − x1 k2 − kx2 − x3 k2 = d2 ,

b2k−2,1 0 b2k−2,3 . . . 0

b2k−2,2k−1  2 2
1 2 1 2
0 b2k−1,2 0 ... b2k−1,2k−2 0 hx2 − x1 , x4 − x3 i = hx2 − x1 , x4 − x1 i − hx2 − x1 , x3 − x1 i = d − d = 0.
2 2
Note that every matrix of this form has determinant zero, because it has k columns spanning a vector This shows that scalar products among vectors which are finite linear combinations of the form
space of dimension at most k − 1.
Call a square matrix of type (C), if it is of the form λ1 x1 + λ2 x2 + · · · + λn xn ,

0 c11 0 c12 . . . 0 c1,k
 where x1 , x2 , . . . , xn are distinct points in S and λ1 , λ2 , . . . , λn are integers with λ1 + λ2 + · · ·+ λn = 0,
 c11 0 c12 0 . . . c1,k 0  are universal across all such sets S in all Hilbert spaces H; in particular, we may conveniently evaluate
them using examples of our choosing, such as the canonical example above in Rn . In fact this property
 
 0 c21 0 c22 . . . 0 c2,k 
trivially follows also when coefficients λi are rational, and hence by continuity any real numbers with
 
C ′ =  c21 0 c22 0 . . . c2,k 0 
 
 .. .. .. .. .. .. ..  sum 0.
 . . . . . . . 
  If S = {x1 , x2 , . . . , xn } is a finite set, we form
 0 ck,1 0 ck,2 . . . 0 ck,k 
ck,1 0 ck,2 0 . . . ck,k 0 1
x= (x1 + x2 + · · · + xn ) ,
n
By permutations of rows and columns, we see that
pick a non-zero vector z ∈ [Span(x1 − x, x2 − x, . . . , xn − x)]⊥ and seek y in the form y = x + λz for
a suitable λ ∈ R. We find that
 
C 0
| det C ′ | = det = | det C|2 ,
0 C
hx1 − y, x2 − yi = hx1 − x − λz, x2 − x − λzi = hx1 − x, x2 − xi + λ2 kzk2 .
where C denotes the k × k-matrix with coefficients ci,j . Therefore, the determinant of any matrix of
hx1 − x, x2 − xi may be computed by our remark above as
type (C) is a perfect square (up to a sign).

Now let X ′ be the matrix obtained from A by replacing the first row by 1 0 0 . . . 0 , and
* ⊤  ⊤ +
d2 1 1 1 1 1 1 1 1
let Y be the matrix obtained from A by replacing the entry a11 by 0. By multi-linearity of the hx1 − x, x2 − xi = − 1, , , . . . , , , − 1, , . . . ,
2 n n n n n n n n
determinant, det(A) = det(X ′ ) + det(Y ). Note that X ′ can be written as Rn
2 2
   
d 2 1 n−2 d
  = −1 + =− .
1 0 2 n n n2 2n
X′ =
v X √
d 2
for some (n − 1) × (n − 1)-matrix X and some column vector v. Then det(A) = det(X) + det(Y ). So the choice λ = √ will make all vectors (xi − y) orthogonal to each other; it is easily
2nkzk d
Now consider two cases. If n is odd, then X is of type (C), and Y is of type (B). Therefore, checked as above that they will also be of length one.
| det(A)| = | det(X)| is a perfect square. If n is even, then X is of type (B), and Y is of type (C); Let now S be an infinite set. Pick an infinite sequence T = {x1 , x2 , . . . , xn , . . .} of distinct points
hence | det(A)| = | det(Y )| is a perfect square. in S. We claim that the sequence
The set of primes can be replaced by any subset of {2} ∪ {3, 5, 7, 9, 11, . . . }.
1
yn = (x1 + x2 + · · · + xn )
n

2 3