20

THE OLYMPIAD CORNER

No. 275

R.E. Woodrow

The year has own by, and it has brought many hanges to Crux and
to the Corner. Of ourse it has been overshadowed by the sudden and un-
timely loss of a great friend and a devoted olleague, Jim Totten, mid-way
through the transition to a new Editor-in-Chief. I think Vazz Linek has done
a wonderful job of stepping in and keeping the journal on tra k with only an
understandable slowing of the produ tion pa e in the interim.
Readers will have noti ed the announ ement at the end of the De em-
ber Corner that Joanne Canape, who has transformed my s ribbles into a
high quality tex le for many years, has de ided that two de ades is enough.
As I ustomarily begin the year by thanking all those who ontributed to the
Corner in the last year, I would be very remiss not to lead o with sin ere
thanks to Joanne.
It is also appropriate to thank those who submitted problem sets for our
use as well as a spe ial thanks to the dedi ated readers who furnish their ni e
solutions whi h we use. Hoping, as always, that I've not missed someone,
here is the list for the 2008 members of the Corner.
Arkady Alt Robert Morewood
Miguel Amengual Covas Andrea Munaro
Jean-Claude Andrieux Vedula N. Murty
Houda Anoun Felix Re io
Ri ardo Barroso Campos Xavier Ros
Mi hel Bataille D.J. Smeenk
Jose Luis Daz-Barrero Babis Stergiou
J. Chris Fisher Daniel Tsai
Kipp Johnson Panos E. Tsaoussoglou
Georey A. Kandall George Tsapakidis
Ioannis Katsikis Jan Verster
R. Laumen Edward T.H. Wang
Salem Maliki Luyan Zhong-Qiao
Pavlos Maragoudakis Li Zhou
Titu Zvonaru

Our apologies to Svetoslav Sav hev for the misspelling of his name in
the De ember 2008 Olympiad.
For your problem solving pleasure in the new year we start o with the
problems of the German Mathemati al Olympiad, Final Round, 2006. My
thanks go to Robert Morewood, Canadian Team Leader to the 47th IMO in
Slovenia 2006, for olle ting them for our use.
 21

German Mathemati al Olympiad

Final Round, Grades 12{13
Muni h, April 29 { May 2, 2006

First Day
1. Determine all positive integers n for whi h the number
zn = 101
| ·{z
· · 101}
2n+1 digits

is a prime.
2. Five points are on the surfa e of a sphere of radius 1. Let amin denote
the smallest distan e (measured along a straight line in spa e) between any
two of these points. What is the maximum value for amin , taken over all
arrangements of the ve points?
3. Find all positive integers n for whi h the numbers 1, 2, 3, . . . , 2n an be
oloured with n olours in su h a way that every olour appears twi e and
every number 1, 2, 3, . . . , n appears exa tly on e as the dieren e of two
numbers with the same olor.
Se ond Day
4. Let D be a point inside the triangle ABC su h that AC − AD ≥ 1 and
BC − BD ≥ 1. Prove that EC − ED ≥ 1 for any point E on the side AB .

5. Let x be a nonzero real number satisfying the equation ax2 + bx + c = 0.

Furthermore, let a, b, and c be integers satisfying |a| + |b| + |c| > 1. Prove
that
1
|x| ≥ .
|a| + |b| + |c| − 1

6. Let a ir le through B and C of a triangle ABC interse t AB and AC in

Y and Z , respe tively. Let P be the interse tion of BZ and CY , and let X
be the interse tion of AP and BC . Let M be the point that is distin t from
X and on the interse tion of the ir um ir le of the triangle XY Z with BC .
Prove that M is the midpoint of BC

Our se ond problem set for this number is a set of sele ted problems
from the Thai Mathemati al Olympiad Examinations 2005. Again, thanks
go to Robert Morewood, team leader to the 47th IMO in Slovenia 2006, for
olle ting them for the Corner.
22

Thai Mathemati al Olympiad Examinations 2005

Sele ted Problems

1. Let P (x), Q(x), and R(x) be polynomials satisfying

1 + x + x2 R(x) .




2xP x3 + Q −x − x2 =

Show that x − 1 is a fa tor of P (x) − Q(x).

2. Find all fun tions f : R → R su h that



f x + y + f (xy) = f f (x + y) + xy

for all x, y ∈ R.
3. Let a, b, and c be positive real numbers. Prove that
3 6
1 + ≥ .
ab + bc + ca a+b+c

4. Let n be a positive integer. Prove that n(n + 1)(n + 2) is not a perfe t

square.
5. Find the least positive integer n su h that 2549| n2545 − 2541 .

6. Do there exist positive integers x, y, and z su h that

2548x + (−2005)y = (−543)z ?

7. Show that there exist positive integers m and n su h that gcd(m, n) = 1

and 2549| ((25 · 49)m + 25n − 2 · 49n ).
8. The median AM of a triangle ABC interse ts its in ir le ω at K and L.
The lines through K and L parallel to BC interse t ω again at X and Y ,
respe tively. The lines AX and AY interse t BC at P and Q. Prove that
BC = CQ. (Shortlist 2005)

9. Let ABC be an a ute-angled triangle with AB 6= AC , let H be its

ortho entre and M the midpoint of BC . Points D on AB and E on AC
are su h that AE = AD and D, H , and E are ollinear. Prove that HM is
orthogonal to the ommon hord of the ir um ir les of triangles ABC and
ADE . (Shortlist 2005)

10. Assume ABC is an isos eles triangle with AB = AC . Suppose that

P is a point on the extension of side BC . X and Y are points on lines AB
and AC su h that P XkAC and P Y kAB . Let T be the midpoint of ar BC .
Prove that P T ⊥ XY . (Iran 2004)
 23

As a third set of problems we give the 46th Ukrainian Mathemati al

Olympiad Final Round 2006 - 11th form problems. Again, thanks go to Robert
Morewood, team leader to the 47th IMO in Slovenia 2006, for olle ting them
for our use.
46th Ukrainian Mathemati al Olympiad 2006
Final Round
th
11 Form

1. (V.V. Plakhotnyk) Prove that for any rational numbers a and b the graph
of the fun tion
f (x) = x3 − 6abx − 2a3 − 4b3 , x ∈ R
has exa tly one point in ommon with the x-axis.
2. (O.A. Sarana) A ir le is divided into 2006 equal ar s by 2006 points.
Baron Mun hausen laims that he an onstru t a losed polygonal urve
with the set of verti es onsisting of these 2006 points su h that amongst its
2006 edges there are no two whi h are parallel to ea h other. Is his laim
true or false?
3. (T.M. Mitelman)
(a) Prove that for any rational number α ∈ (0, 1)  there exists an innite
set of real numbers that satisfy the equation x x{x} = α and any
 

two of them have the same fra tional part. (The fra tional part of a real
number a is given by {a} = a − ⌊a⌋, where ⌊a⌋ is its integer part, that
is, the greatest integer that does not ex eed a.)
(b) Prove that for any rational number α ∈ (0, 1)  there exists an innite
set of real numbers that satisfy the equation x x{x} = α and any
 

two of them have dierent fra tional parts.

4. (V.A. Yasinskiy) Two ir les ω1 and ω2 interse t ea h other at two distin t
points A and B . The tangent line of the ir le ω1 at the point A and the
tangent line of the ir le ω2 at the point B meet at point C . The rst of
these two lines interse ts the ir le ω2 for the se ond time at point T 6= A.
The point X (distin t from A and B ) is on the ir le ω1 , and the line XA
interse ts the ir le ω2 for the se ond time at point Y (distin t from A). The
lines Y B and XC meet at point Z . Prove that T Z is parallel to XY .
5. (O.O. Kur henko) Prove that for any real numbers x and y
| cos x| + | cos y| + | cos(x + y)| ≥ 1 .

6. (T.M. Mitelman) Find all fun tions f : R → R su h that

f x3 + y 3 = x2 f (x) + yf (y 2 )
for all real numbers x and y.
24

7. (V.A. Yasinskiy) A point M lies inside a ube ABCDA1 B1 C1 D1 . Points

A′ , B ′ , C ′ , D ′ , A′1 , B1′ , C1′ , and D1′ belong to the rays M A, M B , M C ,
M D , M A1 , M B1 , BC1 , and M D1 respe tively. Prove that if the poly-
hedron A′ B ′ C ′ D′ A′1 B1′ C1′ D1′ is a parallelepiped (that is, all of its fa es are
parallelograms), then it is a ube.
8. (V.A. Yasinskiy) There are n ≥ 3 soldiers in aptain Petrenko's squad,
no two of the same height. The aptain orders them to stand single-le (not
ne essarily sorted by height). A \wave" is any subsequen e of (not ne essar-
ily next to ea h other) soldiers in this line su h that the rst (leftmost) soldier
in the wave is taller than the se ond soldier in it, but the se ond soldier in it
is shorter than the third one, who is in turn taller than the fourth one, and
so on. (For example, if n = 9, the soldiers are enumerated by height, and
the aptain lines them up as 9, 3, 5, 7, 1, 2, 6, 4, 8 then a longest wave for
this line-up is 9, 3, 7, 1, 6, 4, 8. However, if the aptain lines them up as
1, 2, 3, 4, 5, 6, 7, 8, 9, then a longest wave onsists of (any) one soldier.)
For every n, onsider the number of possible lines with the longest waves of
even lengths and the number of possible lines with the longest waves of odd
lengths. Whi h of these numbers is bigger?

Continuing with problems for readers to solve we give the Cze h-Polish-
Slovak Mathemati s Competition written on June 26-28, 2006 at Z ilina,
Slovakia. Thanks again go to Robert Morewood, Canadian team leader to
the 47th IMO in Slovenia 2006, for olle ting them for our use.
Cze h-Polish-Slovak Mathemati s Competition 2006

1. Five distin t points A, B , C , D, and E lie in this order on a ir le of

radius r and satisfy AC = BD = CE = r. Prove that the ortho entres
of the triangles ACD, BCD, and BCE are the verti es of a right-angled
triangle.
2. There are n hildren sitting at a round table. Erika is the oldest among
them and she has n andies. No other hild has any andy. Erika distributes
the andies as follows. In every round, all the hildren with at least two
andies show their hands. Erika hooses one of them and he/she gives one
andy to ea h of the hildren sitting next to him/her. (So in the rst round
Erika must hoose herself to begin the distribution.) For whi h n ≥ 3 is it
possible to redistribute the andies so that ea h hild has exa tly one andy?
3. The sum of four real numbers is 9 and the sum of their squares is 21.
Prove that these four numbers an be labelled as a, b, c, and d so that the
inequality ab − cd ≥ 2 holds.
 25

4. Prove that for every positive integer k there is a positive integer n su h

that the de imal representation of 2n has a blo k of exa tly k onse utive
zeros, that is, 2n = · · · a00 · · · 0b · · · , where a and b are nonzero digits with
k zeros between them.

n=1 su h that an 6= −1 and

5. Find the number of integer sequen es (an )∞
an + 2006
an+2 =
an+1 + 1

for every positive integer n.

6. Is there a onvex pentagon A1 A2 . . . A5 su h that for ea h i the lines
Ai Ai+3 and Ai+1 Ai+2 interse t in Bi and the points B1 , B2 , . . . , B5 are
ollinear? (By onvention A6 = A1 , A7 = A2 , and A8 = A3 .)

Our nal problem set for this issue is the XXI Olimpiadi Italiano della
Matemati a, Cesenati o, written 5 May 2006. Thanks again go to Robert
Morewood, Canadian team leader to the 47th IMO in Slovenia, for olle ting
them for our use.
XXI Olimpiadi Italiano della Matemati a
Cesenati o
May 5, 2006

1. Rose and Savino play a game with a de k of traditional Neapolitan playing

ards whi h onsists of 40 ards of four dierent suits, numbered 1 to 10. At
the start ea h player has 20 ards. Taking turns, one shows a ard on the
table. Whenever some ards on the table add to exa tly 15, these are then
removed from the game (if the sum 15 an be obtained in more than one way,
the player who last moved de ides whi h ards adding to 15 to remove). At
the end of the game only one ard, a 9, is left on the table. Savino holds two
ards numbered 3 and 5, and Rose holds one ard. What is the number of
Rose's ard?
2. Find all values of m, n, and p su h that
pn + 144 = m2 ,

where m and n are positive integers and p is a prime number.

3. Let A and B be two points on a ir le Γ su h that AB is not a diameter.
Let P be a point on Γ dierent from A and B , and let H be the ortho entre
of the triangle ABP . Find the lo us of H as P varies over all points of Γ
dierent from A and B .
26

4. On an innite hessboard all the .......................................................................................................................................................................................

positive integers are written in as- ... 17 ...... 16 ...... 15 ...... 14 ...... 13 ...
ending order along a spiral, starting ......................................................................................................................................................................
from 1 and pro eeding anti lo kwise;
...... ..... ..... ..... .... ....
..... 18 ..... 5 .... 4 .... 3 ..... 12 ....
a portion of the hessboard is shown
................................................................................................................................
...... ...... ..... ...... ..... .....
in the gure. ..........19 ... 6 ... 1 .. 2 ... 11 .
... ..............................................................................................................................................
A "right half-line" of the hess- ...... 20 ....... 7 ....... 8 ....... 9 ....... 10 ......
board is the set of squares given by .................................................................................................................................................................
a square C and by all squares in the
...... 21 ........ 22 ........ 23 ........ 24 ........ 25 ........
...................................................................................................................................................................................
same row as C and to the right of C .
(a) Prove that there exists a right half-line none of whose squares ontains
a multiple of 3.
(b) Determine if there exist innitely many pairwise disjoint right half-lines
none of whose squares ontains a multiple of 3.

5. Consider the inequality

(x1 + · · · + xn )2 ≥ 4(x1 x2 + x2 x3 + · · · + xn x1 ) .

(a) Determine for whi h n ≥ 3 the inequality holds true for all possible
hoi es of positive real numbers x1 , x2 , . . . , xn .
(b) Determine for whi h n ≥ 3 the inequality holds true for all possible
hoi es of any real numbers x1 , x2 , . . . , xn .

6. Albert and Barbara play a game. At the start there are some piles of oins
on the table, not all ne essarily with the same number of oins. The players
move in turn and Albert starts. At ea h turn a player may either take a oin
from a pile or divide a pile into two piles with ea h pile ontaining at least
one oin (a player may exer ise only one of these options).
The one who takes the last oin wins the game. In terms of the number
of piles and the number of oins in ea h pile at the start, determine whi h of
the players has a winning strategy.

Now we turn to our le of solutions from the readers to problems from

the Mar h 2008 number of the Corner and the Estonian IMO Sele tion Con-
test 2004-2005, given at [2008 : 79-80℄.
3. Find all pairs (x, y) of positive integers satisfying (x + y)x = xy .
Solution by Konstantine Zelator, University of Toledo, Toledo, OH, USA.
We show that there are exa tly two su h pairs, (x, y) = (2, 6), (3, 6).
We will make use of two basi fa ts from elementary number theory.
 27

(a) If a, b ∈ Z and gcd(a, b) = 1 , then gcd(am , bn ) = 1 for any positive

integers m and n.
(b) If a|b, a is a positive integer, b ∈ Z, and gcd(a, b) = 1, then a = 1.
Let x and y be positive integers that satisfy (x + y)x = xy . Write
d = gcd(x, y) and let x = dx1 , y = dy1 , where x1 and y1 are positive
integers that are relatively prime. Substituting for x and y yields
dx1 d · (x1 + y1 )x1 d = x1y1 d · dy1 d . (1)
We make ases by omparing the sizes of x1 and y1 .
Case 1. Suppose that x1 = y1 . Sin e gcd(x1 , y1 ) = 1, we have x1 = y1 = 1.
Thus, equation (1) be omes dd · 2d = dd , whi h is impossible sin e d ≥ 1.
Case 2. Suppose that y1 < x1 . Then x1 − y1 is a positive integer and from
equation (1) we obtain
dd(x1 −y1 ) · (x1 + y1 )x1 d = xy11 d . (2)
Sin e gcd(x1 , y1 ) = 1 it follows that gcd(x1 + y1 , x1 ) = 1. By (a) above, we
have gcd (x1 + y1 )x d , xy1 d = 1. However, by equation (2), the positive
1


1

integer (x1 +y1 )x d is a divisor of xy1 d . Sin e these two integers are relatively
1 1

prime, it follows by (b) that (x1 + y1 )x d = 1, whi h is impossible sin e

1

x1 + y1 ≥ 2 and x1 · d ≥ 1.
Case 3. Suppose that x1 < y1 . From (1) we obtain
(x1 + y1 )x1 d = x1y1 d · dd(y1 −x1 ) . (3)
Sin e gcd(x1 , y1 ) = 1 we have gcd (x1 + y1 )x1 d , xy11 d = 1 and from equa-

tion (3) we see that xy11 d is a divisor of (x1 + y1 )x1 d , whi h implies that
xy11 d = 1. Sin e y1 d is a positive integer this means that x1 = 1. Going ba k
to equation (3) we see that (1 + y1 )d = dd(y1 −1) , hen e
1 + y1 = dy1 −1 . (4)
Note that d 6= 1; otherwise equation (4) be omes 1 + y1 = 1, ontrary to the
fa t that y1 is a positive integer. Thus, d ≥ 2. Sin e y1 = 1 does not satisfy
equation (4), we also have y1 ≥ 2. Setting k = y1 − 1 equation (4) then
be omes k + 2 = dk . By Indu tion (or the Binomial Theorem) we obtain
2k > k + 2 for all integers k ≥ 3. Sin e dk ≥ 2k , it follows from k + 2 = dk
that k = 1 or k = 2.
For k = 2 we have 4 = d2 , hen e d = 2. From 2 = k = y1 − 1 we then
have y1 = 3. Re all that x1 = 1. Going ba k, we have x = x1 d = 1 · 2 = 2
and y = y1 d = 3 · 2 = 6. This is the solution (x, y) = (2, 6).
Similarly, for k = 1 we have d = 3. Then y1 = k + 1 = 2 and sin e
x1 = 1 we obtain x = dx1 = 3 · 1 = 3 and y = dy1 = 3 · 2 = 6. This is the
other solution (x, y) = (3, 6).
28

4. Find all pairs (a, b) of real numbers su h that all roots of the polynomials
6x2 − 24x − 4a and x3 + ax2 + bx − 8 are non-negative real numbers.

Solution by Titu Zvonaru, Comane  sti, Romania.

Let β1 , β2 , and β3 be the roots of the polynomial x3 + ax2 + bx − 8, so
that x3 + ax2 + bx − 8 = (x − β1 )(x − β2 )(x − β3 ). Comparing oeÆ ients
yields β1 + β2 + β3 = −a and β1 β2 β3 = 8. Sin e β1 , β2 , and β3 are
nonnegative real numbers, by the AM-GM Inequality we have
p
β1 + β2 + β3 ≥ 3 3
β1 β2 β3 ,

hen e −a ≥ 6 or a ≤ −6. The equation 6x2 − 24x − 4a = 0 has real roots

if and only if 242 − 4 · (−4a) · 6 ≥ 0, whi h implies 242 + 24 · 4a ≥ 0 and
hen e a ≥ −6. Therefore, a = −6. √
Now we have β1 + β2 + β3 = 6 = 3 β1 β2 β3 , from whi h it follows
3

that β1 = β2 = β3 = 2 and b = 12. Thus, the only pair satisfying the

ondition is (a, b) = (−6, 12).

Next we turn to a solution to a problem of the Trentieme Olympiad

Mathematique
 Belge Maxi Finale, Mer redi 20 avril 2005 given at [2008 : 80℄.
3. Dans le triangle ABC , les droites AE et CD sont les bisse tri es
interieures
 des angles ∠BAC et ∠ACB respe tivement ; E appartient a BC
et D appartient a AB . Pour quelles amplitudes de l'angle ∠ABC a-t-on
ertainement
(a) |AD| + |EC| = |AC|? (b) |AD| + |EC| > |AC|?
( ) |AD| + |EC| < |AC|?
Solution by Titu Zvonaru, Comane
 sti, Romania
As usual let a = BC , b = CA, A
and c = AB . The rst equation below ...

is the Angle Bise tor Theorem; the fol-

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

lowing equations are equivalent to it:

.. ....
.. .... ....
.. ... ....
.. ....
.. ...
..
. .. ....
....
. ...
... ... ....
. ....
BE AB
.

;
. ... ....
... ..
= . ... ....
.. ....
... ... ....
EC AC D ...
...........
..
...
...
....
....
....
. .........
. .. ....
BE c
;
. .......... ....
... .......... ... ....
= .. .......... ... ....
. ........... ....
... ..........
EC b . . . ... .
. ....
.. ... ........... . ....
... ... ..
............ ....
....
.. . .
....
..........
. ... ....
BE + EC b+c
;
... .. ....
.......... ....
=
.
. .
.. ...
............ ....
. ....
... ... .......... .
EC b
..........
.
.
. ...
.. .......... .......
... ... .......... ....
........ ...
...
..............................................................................................................................................................................................................................
ab
EC =
b+c
. B E C

bc
Similarly, AD = a+b
.
 29

We have
bc ab
|AD| + |EC| − |AC| = + − b
a+b b+c
 
c(b + c) + a(a + b) − (a + b)(b + c)
= b .
(a + b)(b + c)

By the Law of Cosines a2 + c2 − b2 = 2ac cos B , so the equation above an

be rewritten as
1


2abc cos(B) −
|AD| + |EC| − |AC| = 2
.
(a + b)(b + c)
Hen e,
|AD| + |EC| = |AC| ⇐⇒ ∠ABC = 60◦ ,
|AD| + |EC| > |AC| ⇐⇒ ∠ABC < 60◦ ,
|AD| + |EC| < |AC| ⇐⇒ ∠ABC > 60◦ .

Next we turn to solutions from our readers to problems of the 2005

Vietnam Mathemati al Olympiad given at [2008 : 81℄.
1. Find the smallest and largest values of√the expression
√ P = x + y , where
x and y are real numbers satisfying x − 3 x + 1 = 3 y + 2 − y .

Solved by Arkady Alt, San Jose, CA, USA; and Mi hel Bataille, Rouen, Fran e.
We give the solution of Bataille.
√
√
We show that Pmin = 21 9 + 3 21 √

and Pmax = 9 + 3 15.
First, x = −1 and y = 12 11 + 3 21 satisfy the onstraint equation
√ √ √ √ √
2
x−3 x + 1 = 3 y + 2−y (easily he ked using 10+2 21 = 3+ 7 )
√ √ √

and P = 12 9+3 21√. Similarly, for x = 12 10+3 15 , y = 21 8+3 15 ,

we have P = 9 + 3 15 and the onstraint√is satised.√ Now, let x and y

satisfy the onstraint equation. Then P = 3 x + 1 + 3 y + 2, so that
 p 
P 2 = 9 P + 3 + 2 (x + 1)(y + 2) . (1)
It follows that P ≥ 0 and P 2 − 9P − 27 ≥ 0. Thus, P is not less than
the positive solution
√
of the quadrati x − 9x − 27 = 0 and we dedu e that
2

P ≥ 2 9 + 3 21 . From the AM-GM Inequality and (1), we obtain

1

P 2 ≤ 9(P + 3 + x + 1 + y + 2) = 9(2P + 6) = 18P + 54 ,

√
orP 2 − 18P − 54 ≤ 0, whi h implies that P ≤ 9 + 3 15. The proof is
omplete.
30

4. Find all
real-valued fun tions f dened on R that satisfy the identity
f f (x − y) = f (x)f (y) − f (x) + f (y) − xy .

Solved by George Apostolopoulos, Messolonghi, Gree e; Mi hel Bataille,

Rouen, Fran e; and Daniel Tsai, student, Taipei Ameri an S hool, Taipei,
Taiwan. We give the write up of Bataille.
It is readily he ked that the fun tion f (x) = −x for all x is a solution.
We show that there are no other solutions. Let f satisfy
(1)


f f (x − y) = f (x)f (y) − f (x) + f (y) − xy
for all x, y ∈ R and let a = f (0). Taking x = y =0 in (1) gives f (a) = a2
and then taking y = x shows that
f (x)2 = x2 + a2
for all real
√ numbers
√ x. In parti ular f (a)
√ = 2a , that is, a = 2a , hen e
2 2 4 2

a ∈ {0,√ 2, − 2}. Assume √ that a = 2. Then for any given x we have

f (x) = x2 + 2 or f (x) = − x2 + 2 and taking y = 0 in (1), we obtain
√ √
2. (2)



f f (x) = 2 − 1 f (x) +
√
√

√
Now, f 2 = f (a) = a2 = 2 so that f f 2 = f (2) = ± 6, but then
√ √
taking x = 2 in (2) yields a ontradi tion sin e 6 6= 3 2 − 2 . Similarly,

2
√
the assumption a = − 2 leads to a ontradi tion. It follows that a = 0,
hen e f (x) = x or f (x) = −x for any given
x. However, if f (x0 ) = x0 for
some nonzero real
number x0 , then
f f (x0 ) = f (x0 ) = x0 , while from (1)
we have f f (x0 ) = f f (x0 − 0) = −f (x0 ) = −x0 . This is impossible
sin e x0 6= 0, hen e f (x) = −x for all real x.
x! + y!
5. Find all triples of non-negative integers (x, y, n) su h that n!
= 3n
(with the onvention 0! = 1).
Solved by Daniel Tsai, student, Taipei Ameri an S hool, Taipei, Taiwan; and
Konstantine Zelator, University of Toledo, Toledo, OH, USA. We give the
solution of Tsai.
Let S be the set of all ordered triples (x, y, n) of nonnegative integers
su h that x! n!
+ y!
= 3n , or equivalently x! + y! = 3n n!. For n = 0, it is
seen at on e that there is no orresponding (x, y, n) ∈ S . For n = 1, let
(x, y, n) ∈ S ; then if x ≥ 3 or y ≥ 3 we have x! + y! ≥ 7 > 3 = 31 1!,
thus x, y < 3 and simple he king yields that {(0, 2, 1), (1, 2, 1), (2, 0, 1),
(2, 1, 1)} ⊂ S .

Lemma Let (x, y, n) ∈ S with n ≥ 2. Then x, y ≥ n and x > n or y > n.

Proof: If x, y ≤ n, then x! + y! ≤ 2n! < 3n n!, so x > n or y > n. If
x! + y! x! y! x!
x > n and y < n, then = + = 3n , a ontradi tion sin e is
n! n! n! n!
 31

y!
an integer but n! is not. Thus, if x > n, then y ≥ n. Similarly, in the ase
y > n we have x ≥ n by symmetry.

We shall prove that for n ≥ 2 there is no orresponding (x, y, n) ∈S

by onsidering ases on n modulo 3.
Case 1. n ≡ 0 (mod 3). Let (x, y, n) ∈ S and assume without loss of
generality that x ≤ y. By the Lemma, n ≤ x ≤ y and one of these two
inequalities is stri t. If x > n, then from x! n!
+ y!
= 3n it follows that
(n+1)|3n . However, n+1 has a prime divisor other than 3, a ontradi tion.
Therefore, n = x < y, and onsequently
x! + y! x! y!
= + = 1 + (n + 1)(n + 2) · · · y = 3n .
n! n! n!
Thus, 3 divides 1+(n+1)(n+2) · · · y and (n+1)(n+2) · · · y ≡ 2 (mod 3),
whi h implies that y = n + 2. However, 1 + (n + 1)(n + 2) < 3n for n ≥ 3
(by indu tion) and 1 + (2 + 1)(2 + 2) 6= 32 , ontradi ting the fa t that
(x, y, n) ∈ S .
Case 2. n ≡ 1 (mod 3). Let (x, y, n) ∈ S and assume without loss of
generality that x ≤ y. By reasoning similar to that of Case 1 it follows that
x = n and y = n + 1. However, 1 + (n + 1) < 3n for ea h integer n ≥ 2,
ontradi ting the fa t that (x, y, n) ∈ S .
Case 3. n ≡ 2 (mod 3). Let (x, y, n) ∈ S and assume without loss of
generality that x ≤ y. By the Lemma, n ≤ x ≤ y and one of these two
inequalities is stri t. If x ≥ n + 2, then from x! n!
+ y!
= 3n it follows that
(n + 2)|3n . However, n + 2 has a prime divisor other than 3, a ontradi tion,
hen e n ≤ x < n + 2.
If x = n, then n = x < y and
x! + y! x! y!
= + = 1 + (n + 1)(n + 2) · · · y = 3n ,
n! n! n!
ontradi ting n + 1 ≡ 0 (mod 3).
If x = n + 1, then
x! + y! x! y!
= + = (n + 1) + (n + 1)(n + 2) · · · y = 3n .
n! n! n!
If furthermore y = n + 1, then
2(n + 1) = (n + 1) + (n + 1)(n + 2) · · · y = 3n

is even, a ontradi tion. Thus, y > n+1 and (n+1) 1+(n+2) . . . y = 3n .

It follows that 1 + (n + 2) · · · y ≡ 0 (mod 3) and (n + 2) · · · y ≡ 2 (mod

3),
whi h implies that y = n + 3. However, (n + 1) 1 + (n + 2)(n + 3) 6= 3n
32

for 1 ≤ n ≤ 5 and by indu tion (n + 1) 1 + (n + 2)(n + 3) for n ≥ 6,

< 3n
ontradi ting the fa t that (x, y, n) ∈ S .
6. Let the sequen e x1 , x2 , x3 , . . . , be dened by x1 = a, where a is a real
number, and the re ursion xn+1 = 3x3n − 7x2n + 5xn for n ≥ 1.
Find all values of a for whi h the sequen e has a nite limit as n tends
to innity, and nd this limit.
Solved by Arkady Alt, San Jose, CA, USA; Mi hel Bataille, Rouen, Fran e;
and Daniel Tsai, student, Taipei Ameri an S hool, Taipei, Taiwan. We give
Bataille's write-up.
Let f (x) = 3x3 − 7x2 + 5x and g(x) = f (x) − x = x(x − 1)(3x − 4).
The sequen e {xn }, whi h satises xn+1 = f (xn ) for all positive integers n,
an only onverge to a root of g(x) = 0. Thus, the only possible nite limits
of {xn } are 0, 1, and 43 . We show that the sequen e is onvergent if and only
if 0 ≤ a ≤ 43 , in whi h ase the limit is 1 ex ept if a = 0 and n→∞lim xn = 0 or
if a = 3 and n→∞
4
lim xn = 3 . 4

Suppose rst a < 0. Sin e g(x) < 0 when x < 0, it follows that
xn < x1 = a < 0 for all positive integers n. If {xn } had a nite limit, ℓ,
we would have ℓ ≤ a, ontradi ting the fa t that ℓ ∈ {0, 1, 43 }. Thus, {xn }
is divergent when a < 0. Using the fa t that g(x) > 0 for x > 43 , similar
reasoning shows that {xn } is divergent when a > 43 .
If a ∈ {0, 1, 43 }, then the sequen e {xn } is onstant.
If a ∈ (1, 34 ), then using f (x) − 1 = (x − 1)2 (3x − 1) an easy indu -
tion shows that 1 < xn+1 < xn for all positive integers n. Thus, {xn } is
de reasing and bounded, hen e onvergent. Its limit ℓ satises ℓ ≥ 1 and
ℓ ∈ {0, 1, 43 }, that is, ℓ = 1.
If a ∈
[ 13 , 1) then x2 = f (a) ≥ 1 and x2 < 43 , as the maximum of f on
[0, 1] is f 59 = 275 243
< 43 . From the previous ase, we see that lim xn = 1.
n→∞
1 1
It remains to study the ase a ∈ 0, 13 . Then, 3m+1 ≤ a < m for


3
some unique positive integer m. If any of the numbers x2 , x3 , . . . , xm is not
less than 13 , let xk be the one with the smallest index. Then 13 ≤ xk < 43 and
by the previous ases {xn }n≥k onverges to 1 and n→∞ lim xn = 1. Otherwise,
noting that f (x) − 3x = x(x − 2)(3x − 1) is positive for x ∈ 0, 13 , we have

1
x2 = f (x1 ) > 3x1 = 3a ≥ ,
3m
1
x3 = f (x2 ) > 3x2 ≥ ,
3m−1
···
1
xm = f (xm−1 ) > 3xm−1 ≥ ,
32
and nally xm+1 > 13 . So {xn }n≥m+1 onverges to 1 and again n→∞
lim xn = 1.
 33

To nish this number of the Corner we give solutions from the readers
to problems of the 2005 German Mathemati al Olympiad, given at [2008 : 82℄.

1. Determine all pairs (x, y) of reals, whi h satisfy the system of equations

x3 + 1 − xy 2 − y 2 = 0,
y 3 + 1 − x2 y − x2 = 0.

Solved by George Apostolopoulos, Messolonghi, Gree e; Konstantine Zela-

tor, University of Toledo, Toledo, OH, USA; and Titu Zvonaru, Comane
 sti,
Romania. We give the write-up of Apostolopoulos.
We subtra t the two equations of the system to obtain

+ xy(x − y) + x2 − y 2 = 0 ,


x3 − y 3

whi h upon fa toring be omes

(x − y)(x + y)(x + y + 1) = 0 .

Thus, y = x or y = −x or y = −x − 1.
If y = x, then x = ±1, hen e (x, y) = (1, 1) or (x, y) = (−1, −1).
If y = −x, then again x = ±1, hen e (x, y) = (1, −1) or (x, y) = (−1, 1).
If y = −x − 1 we substitute into the rst equation to obtain

x3 + 1 − x(x + 1)2 − (x + 1)2 = −3x(x + 1) = 0 ,

hen e x = 0 or x = −1 and (x, y) = (0, −1) or (x, y) = (−1, 0).

2. Let A, B , and C be three distin t points on the ir le k. Let the lines h

and g ea h be perpendi ular to BC with h passing through B and g passing
through C . The perpendi ular bise tor of AB meets h in F and the perpen-
di ular bise tor of AC meets g in G. Prove that the produ t |BF | · |CG| is
independent of the hoi e of A, whenever B and C are xed.
Solved by Miguel Amengual Covas, Cala Figuera, Mallor a, Spain; Mi hel
Bataille, Rouen, Fran e; Konstantine Zelator, University of Toledo, Toledo,
OH, USA; and Titu Zvonaru, Comane
 sti, Romania. We give the solution of
Amengual Covas.
Let A′ be the point diametri ally opposite to A. Let M and N be
the midpoints of the segments AB and AC , respe tively. Let B ′ and C ′
be the se ond points of interse tion of the lines g and h with the ir le k,
respe tively.
34
. .
.... ....
... ....
.... A ....
h
....
...
.... .... ............. .. ......... .. ..... . .
.
.
.
.
.
.
.
.... . .. .. . .
..
...
.
..
..
..... ... .
................................................... g ....
...
..............
.. .
.. .
...... ... .... .... ............. ............... ...
.... ......
...... .. .
... ........ ...
.... ............ . ....... . .. ... ... ........
............ ........... ....
... ..
........ ..
... ..
....... . . . .... ... ...
......... ..
.. ....
.... . ..
. .... ............. .
. . ... ... .. .
. ........ ....
......
.. ..
. ... ... ... ........ ..... ....
.... ........ .................. . .. . .
′ ... .... ........... ′
. . . ... ........ ... ..
C
. . .
B
. .. . ..
.. .. .
. ......... ... ...
.
.. .... ... . ..........................
............... .. .

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

.........
......... ... ... ... ...... .........
................ .. .. ..............
..... ..... .... ............ . ... ..
.
.
..
.. ................. ..... .....
.. .. .. ...... .. ... .. ..
.. ................ ............ ... ... ....... .. ... ..
... ...
... ...... .... .... ....
... . ..... ....... ... .......
F .. ............. ..
.. ..... ................. .... ... .. .. ..
.......................
...
G
..
... ..
. ... ......... ............ ... .
. .
. .... .. . ..
.. .... .. ........... ............ .
... ....... ......... .
....
..
..
... ....
... ........ ....... .
. .......... .... ...................... .
..
...
............
. . ..

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

.........
... ... ... .
.
. ..... . ...
.......... .
N
... . ........ .
. .... . ... .. . .
. ...
M
... .... .. .. . ........ .
. ..... . . .
.
.... .... ... ... ...
.
.
.
. .......... ... ... .
.
.
. ...
... ... ....... . ...... .... .. ...
... .
. ....... .. ....... .
.... .
. ..
k
.... ..... ... ... .............. .
. .
.
. ..
... .... ...... .
. . . . . ............. ... .... .
.
.
. ...
... ..
. . ...... .......... .... ........... .. . ... ..
.
. ..
.
.. .
. .. . .. ........
. .
.
.
.......
. . ... ... .
.
. ..
.. . . ... .. ....... ... . .
.. .... ... .... ....... ... .......
....... ....
.. .... ...
.
..
.. ... ... ... ............ ... ........
...
... ... ..
... .... .. ... ......... .. .... ..
.. .
. ... ......
... ..
.
. . .............. ...
. .
. .
.
..
.. . . ... . .. ....... ... ..
.. ..... ... ....... .... ... ... ....... ... .... ..
..
... . .. ....... ... ... ... ....... . .. ..
... .......... .... ............ ... .. .. .......
....... ............... ....
... .... ...... .........
. .. ...
.
. .. .. ..... ...... .. .
. ..
... ... .. ....... .. ....... ... ... ..
... .. .. ....... ... ... ... ....... .... ...
............... . ... .. ........
.....................................................................................................................................................................................................................................................
............ ... .. . . ..........
B
.
C
.... ......... . .. ....
.... ......... ... .
.
.
.. ... .. ..... . ...... ......
.
..
.... ........ ... .. .. ......... ........
..... .
....... ................ ... .. .. ......... .
... ..... .... ......... ......
....... .
........ ................ ......... .......
......
..
...

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

.....

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

A′

Sin e C ′ BC and BCB ′ are right angles, the segments BB ′ and CC ′

are diameters of the ir le k. Thus, the two quadrilaterals AC ′ A′ C and
ABA′ B ′ are re tangles, and hen e parallelograms, so we have

|C ′ A′ | = |CA| and |A′ B ′ | = |AB| .

Sin e the right triangles BM F and A′ C ′ A are similar, as are the right tri-
angles CN G and A′ B ′ A, we have
1
BF BM 2
AB
′
= =
AA C ′ A′ CA
and 1
CG CN CA
′
= = 2
,
AA A′ B ′ AB
from whi h we obtain
1
BF · CG = |AA′ |2
4
as the square of the radius of k, whi h is independent of the hoi e of A.
3. A lamp is pla ed at ea h latti e point (x, y) in the plane (that is, x and
y are both integers). At time t = 0 exa tly one lamp is swit hed on. At
any integer time t ≥ 1, exa tly those lamps are swit hed on whi h are at a
distan e of 2005 from some lamp whi h is already swit hed on. Prove that
every lamp will be swit hed on at some time.
Solution by Titu Zvonaru, Comane
 sti, Romania.
Assume
√ that at time t = 0 the lamp at O(0, 0) is swit hed on. Sin e
2005 = 13572 + 14762 then at some time the lamps at the following latti e
 35

points will be swit hed on:

A1 (1357, 1476) , O1 (2 · 1357, 0) ,
A2 (3 · 1357, 1476) , O2 (4 · 1357, 0) ,
.. ..
. .
Ak (2k − 1) · 1357, 1476 , Ok (2k · 1357, 0) ;

and then the lamps at these latti e points will be swit hed on:
B1 (2k · 1357 − 2005, 0) ,
B2 (2k · 1357 − 2 · 2005, 0) ,
B3 (2k · 1357 − 3 · 2005, 0) ,
..
.
Bt (2k · 1357 − t · 2005, 0) .
The equation 2k · 1357 − 2005t = 1 is the same as 2714k − 2005t = 1,
whi h has a solution in positive integers k, t be ause gcd(2714, 2005) = 1,
for example, 2714 · 1134 − 2005 · 1535 = 1. Thus the lamp at (1, 0) will be
swit hed on at some time. It follows (by symmetry) that every lamp will be
swit hed on at some time.
4. LetQ(n) denote thesum of the digits of the positive integer n. Prove
that Q Q Q(20052005 ) = 7.

Solution by Titu Zvonaru, Comane

 sti, Romania.


It is well known that Q(n) ≡ n (mod 9). Let k = QQ Q(20052005 ) ,
then k ≡ 20052005 (mod 9) and we have

668
20052005 ≡ (−2)2005 = −2 · 22004 = −2 · 23
≡ −2(−1)668 = −2 ≡ 7 (mod 9) ,
so that k ≡ 7 (mod 9).
The number 20052005 has at most 4 · 2005 = 8020 digits. Hen e,
is at most 9 · 8020 = 72180. This implies that
2005


Q 2005


Q Q 20052005 ≤ 5 · 9 = 45
 
and hen e k = Q Q Q(20052005 ) is at most Q(39) = 12.

Altogether, k is an integer satisfying 0 ≤ k ≤ 12 and k ≡ 7 (mod 9),

hen e k = 7, as desired.

That ompletes the Corner for this issue. Send me your ni e solutions,
generalizations, and Olympiad problem sets.