XII Mexedyxapodnas Kaymuxosenaa oausmuada no Mamemamue
Aaa, 2017
15 anpapa 2017 roga, 9.00-13.30
Bropoii zenn
‘Karka satava onenmpaeres 8 7 6as.108)
4, Heppnte & tenon ay, a, ..., a4 HocaexonarestsHocTH (ay) ~ pascumiisie natypasty-
ble HEA, a TPH N > k GHCHO dy ~ HANMeHbIIee HaTypAMIbHOE “HCO, He MpeneTABHMoe
1B DIKE CYMMBI HecKOsIBKUX (nOIMOAKHO, OAHOTO) H3 HCE ay, M2, .-.5 Gyr. DoKarkure,
TO dy = 2dy—1 TPH BEX FOCTATOUHO GONIbIIHX n.
5. Ina Kaxksoro HatypanbHioro k oGosnaum sepes C(k) cymay acex pascii4Hbix
npocroix aesuresteit ncaa k. Hanpumep, C(1) = 0, C(2) = 2, C(45) = 8. Hafianre nce
narypactonne n, 219 Koropnx C(2" +1) = C(n).
6. B npocrpancrse ambi npasiwubit verpasnp ABCD 1 npousnospubie vowKH M
V. Toxasxwre nepasencrso
"
MA-NA+MB-NB+MC-NC>MD-ND.
(Terpagap naasinaeres npaouannnst, ecstn nce met» ero peep panne.)
XII International Zhautykou Olympiad in Mathematics
Almaty, 2017
January 15, 9.00-13.30
Second day
(Bach problem is worth 7 points)
4, Initial terms a1, a2, ..., ay of a sequence (aq) are different positive integers, and
for n > k the number a, is the minimum positive integer not representable as a sum of
some of the numbers a1, 42, ..., dn—1 (maybe one of them). Prove that ay = 2ay—1 for
all large enough n.
5. Let C(k) denotes the sum of all different prime divisors of a positive integer k.
For example, C(1) = 0, C(2) = 2, C(45) = 8. Find all positive integers n such that
C(2" +1) =C(n).
6. A regular tetrahedron ABCD and points M, N are given in space. Prove the
inequality
MA-NA+MB-NB+MC-NC>MD-ND.
(A tetrahedron is called regular if all its six edges are equal.)4. Initial terms a1, a3, ..., ay of a sequence (ay) are different. positive integers, and for n > k the
umber a, is the minimum positive integer not representable as a sum of some of the numbers a1, a2, ..»
1-1 (maybe one of them). Prove that. a, = 2¢,- for all large enough
Solution. For each n > k we consider the set of all positive integers not exceeding a +2 +... dy
the numbers in this set that are not sums of elements of the set {a),02,....dq} We call gaps. If the set of
gaps is not empty, then d,.41 is the minimum gap, otherwise dn41 = a1 +2 +...+ a, +1. We claim that
when n increases the number of gaps decreases.
Note that if # <8 = a) +42 +...+ dq is a sum of several numbers in the set {ay,42,-.-.4y}, then
— Lis also such a sum: itis the sum of all a; not included in the original sum.
The fact that dq4 is the minimum gap means that all the numbers from 1 to dng — 1 are sums of
some of the numbers a1, a2, .-. dq. ‘Therefore all the numbers from a1 +... dy ~G@n41 10 a1 +--+ dy
are such sums too. Adding aaa we see that all the numbers from ay +...+4q tO.) 4.0. @q + @n4i are
sums of fome of the numbers ay, a, ..., dn41- ‘Thus, when n increases by 1, no new gaps occur, and at
least one old gap (ay41 itself) disappears, q.e.d
Therefore at some moment there will be no gaps.
We see that dni = a1 +++. y+ 1 for all large enough n. Then dngy = a1 +--+ ty + dyer $1 =
(0) $b ay + 1) Fant = Ung, Qed.
5. Let C(A) denotes the sum of all different prime divisors of a positive integer k. For example, C(1) = 0.
C(U5) = 8. Find all positive integers n such that C(2" + 1) = C(n)
‘The answer is 0
Solution. Let P(t) be the largest prime divisor of a positive integer > 1
Let m be the largest odd divisor of n: n= 2m. Then 2"4+1=2?™41=a™ 41, where a
& > 0, that is, n is even, then C(n) = C(m) +2 and C(2" +1) =C(a” +1)
We need the following two lemmas
Lemma 1. For every prime p > 2 we have P( “t )
Proof, Let P (
(a? — Lat! 1) 1") — 1. The greatest common divisor (2p.q— 1) is even and must equal 2p or
2. In the first case 2p divides q— 1, whence q > 2p-+ 1. In the second case q divides a! — 1 but not a—1
(because a? +1 is divisible by q), that is, @=—1 (mod q). Then At p. (mod q) and
p=4.
Lemma 2. If p; and p» are different odd primes then P(t it) #P( wt )
Proof. ine (se 41) = p(
al2P2ro) —
por p(s} i) > Qp+l
) = 4. It follows from Fermat's little theorem that q divides 27-1 ~ 1 and therefore
1
Z241) = g then q divides a2?'—1 and a®P—1, therefore (a2?!—1, a2
a? — 1 and hence a+ 1, but then pj =q and p, = q, a contradiction
We are ready now to slve the problem. Let pi, ..., Ps be all the prime divisors of n. It follows from
Lemma ? that be 1
19” p (a+ p (ait
> we
ce’ +y>r(S4")4 +0(S 4")
WCQ" +1) > P(E) bP
LAIS., that is,
-I)=
(253), then 2" 41 has at least one prime divisor not summed in the
sg" > (arith > (arti >, 23> Cln
ce +yor(S A) ya e(S et) a> Pitee tps $3 >C(n)
Lherefore we can assume the equality
Ce" += (Ss) fet (os)
an ésuch that P (2351) # p,, then C2"41) > pi t-.-t ps tpi +1 > Cla)
It remains to consider the case when P (S44) pi for all i, In this case we have C(n) = C(2"+1) =
Pit...+ps, 80 must be odd and a But 2? =2 (mod p) for all odd prime p, therefore p > 3 cannot
divide 2? + 1. Thus s with some positive integral r. The number 2” +1 = 2°" +1 must
be a power of 3. However 19 divides this number for r = 2 and consequently for all r > 2. Thus the only
remaining case is n = 3, which obviously satisfies the condition.
If in this case ther6. A regular tetrahedron ABCD and points M, N are given in space. Prove the inequality
MA-NA+MB-NB4+MC-NC > MD-ND.
(A tetrahedron is called regular if all its six edges are equal.)
Solution. We need the following
Lemma 1. For every different points 4, B,
J, D the inequality
AB-CD+BC-AD > AC. BD
holds.
Proof. Consider the point 4 on the ray DA such that Dy = zfq. In the same way we take the points
B, and C; on the rays DB and DC. Since 94} = 94 = = lpg, it follows from Esealae of the triangles
DAB and DByAy that Ay By 5. Similarly, By pine and CiAy = p&4pz (1). Substituting
these equalities in the triangle inequality Ay By + ByCy > AC) we obtain AB-CD + BC-AD > AC BD.
Lemma 2. Kor every points iM, \ in the plane of the triangle ABC
CM-CN :
caorn 7h (
Proof. In the plane ABU we consider the point K such that ZABM = ZK BU, ZMAB = ZCKB.
Note that
Chee eee Q
BK AB'BK BC'BK AB" :
Applying lemma 1 to the points A, N, C, K we have AN-CK +CN-AK > AC-NK. Triangle inequality
NK > BK — BN gives us AN -CK-+CN- AK > AC -(BK ~ BN). Hence we obtain
AN-CK
RK TAC
K
"BR
(3)
It follows from (3) and (2) that 44-4 4 BabA 4 GALEN > 1
Corollary. The inequality (*) remains true when one of the points M, N, or both, lie outside the
plane of the triangle ABC.
It follows from lemma 2 when instead of M and NV it is applied to their projections onto the plane of
the triangle ABC,
We are ready now to solve the problem. On the ray DA we consider the point A, such that DAy =
Ina similar way we take points By, Ci, Mr, 1 on the rays DB, DC, DM, DN
Applying the corollary of Lemma 2 to the points My, Vy and the triangle A, B,C we get the inequality
Ay My» AyNy + ByMy + ByNy + Ci My ~Ci Ni > AyB}; using equations similar to (1) we obtain
Da
AM AN __BM BN, _CM CN >( AB )
DA-DM DA-DN * DB-DM DB-DN * DC-DM DC“ DN * \DA-DB
whence
AM+AN + BM+BN+CM-CN > DM-DN
aed