4th Benelux Mathematical Olympiad

20–22 April 2012 — Namur, Belgium

Solutions

Problem 1. A sequence a1 , a2 , . . . , an , . . . of natural numbers is defined by the rule

an+1 = an + bn (n = 1, 2, . . . )

where bn is the last digit of an . Prove that such a sequence contains infinitely many powers of
2 if and only if a1 is not divisible by 5.

Solution. First we can observe that:

• If a1 is divisible by 5, then an = a2 = 0 (mod 10) ∀n ≥ 2.

• If a1 is not divisible by 5, then for n ≥ 2: an is even, the sequence bn is periodic, its

period is a cyclic permutation of (2, 4, 8, 6), and an+4 = an + 20.

(a) Let us suppose that a1 is divisible by 5.

Since 2k 6= 0 (mod 10) for any k ∈ N, the sequence does not contain any power of 2 for
n ≥ 2.

(b) Let us suppose that a1 is not divisible by 5.

We can remark that the sequence of powers of 2 modulo 20 respects the period (12, 4, 8, 16)
starting with 25 = 32. We choose j such that aj = 2 (mod 10) (i.e. bj = 2) and look at
the parity of its penultimate digit.

• If aj = 12 (mod 20), then the numbers aj+4k , k ∈ N, represent all the numbers
congruent to 12 (mod 20) and greater than aj , so all powers of 2 congruent to 12
(mod 20) and greater than aj appear in the sequence.
• If aj = 2 (mod 20), then the numbers aj+1+4k , k ∈ N, represent all the numbers
congruent to 4 (mod 20) and greater than aj+1 , so all powers of 2 congruent to 4
(mod 20) and greater than aj+1 appear in the sequence.

Thus, the sequence contains infinitely many powers of 2.

Alternative 1 for (b). We choose j such that aj = 2 (mod 10) (i.e. bj = 2).

• If aj = 20t + 12 for some t ∈ N, then aj+4k = aj + 20k = 20(t + k) + 12, ∀k ∈ N. We

4s+3
obtain infinitely many powers of 2 by taking k = 2 5 −3 − t (with s ∈ N large enough to
have k > 0) since 24s+3 = 3 (mod 5), ∀s ∈ N.

1
 • If aj = 20t + 2 for some t ∈ N, then aj+1+4k = aj+1 + 20k = 20(t + k) + 4, ∀k ∈ N. We
4s
obtain infinitely many powers of 2 by taking k = 2 5−1 − t (with s ∈ N large enough to
have k > 0) since 24s = 1 (mod 5), ∀s ∈ N.

Alternative 2 for (b). Choose j such that aj is a multiple of 4, i.e. aj = 4q (such a j always
exists since an+1 = an + 2 for infinitely many n). Then we have aj+4k = aj + 20k = 4(q + 5k).
Let us look for (k, m) such that

aj+4k = 2m ⇐⇒ 4(q + 5k) = 2m ⇐⇒ q + 5k = 2m−2 ⇐⇒ 2m−2 = q (mod 5).

Since q could not be a multiple of 5, we have q ∈ {1, 2, 3, 4} (mod 5). Since the sequence
2m−2 (mod 5) is periodic with period (1, 2, 4, 3), we find that 2m−2 = q (mod 5) happens for
infinitely many values of m. Hence 2m−2 = q + 5k is solvable for infinitely many pairs (k, m).
Noting that m determines k and that k is nonnegative as soon as m is large enough concludes
the proof.

Alternative 3 for (b). We shall show that for any n > 1 there is some k ≥ n such that ak
is a power of 2. First, we observe that we can always find m ∈ {n, n + 1, n + 2, n + 3} such
that am is divisible by 4. If am is not a power of 2, we write am = 2b c with b ≥ 2 and c > 1
odd. Then we have
c+5
am+4·2b−2 = am + 20 (2b−2 ) = 2b c + 5 · 2b = 2b+1 ·
2
If c > 5, we have c+5
2
< c and hence the odd factor of am+4.2b−2 is strictly smaller than the odd
0
factor of am . Therefore there is some m0 > m such that am0 = 2b c0 with c0 odd and ≤ 5. The
0
case c0 = 5 is forbidden. If c0 = 1, then am0 is a power of 2. If c0 = 3, then am0 +4.2b0 −2 = 2b +3 is
a power of 2.

2
Problem 2. Find all quadruples (a, b, c, d) of positive real numbers such that abcd = 1,
a2012 + 2012b = 2012c + d2012 and 2012a + b2012 = c2012 + 2012d.

Solution. Rewrite the last two equations into

a2012 − d2012 = 2012(c − b) and c2012 − b2012 = 2012(a − d) (1)

and observe that a = d holds if and only if c = b holds. In that case, the last two equa-
tions are satisfied, and condition abcd = 1 leads to a set of valid quadruples of the form
(a, b, c, d) = (t, 1t , 1t , t) for any t > 0.
We show that there are no other solutions. Assume that a 6= d and c 6= b. Multiply both sides
of (1) to obtain
(a2012 − d2012 )(c2012 − b2012 ) = 20122 (c − b)(a − d)
and divide the left-hand side by the (nonzero) right-hand side to get

a2011 + · · · + a2011−i di + · · · + d2011 c2011 + · · · + c2011−i bi + · · · + b2011

· =1.
2012 2012
Now apply the arithmetic-geometric mean inequality to the first factor

a2011 + · · · + a2011−i di + · · · + d2011

q
2012 2011×2012 2011
> (ad) 2 = (ad) 2 .
2012
The inequality is strict, since equality holds only if all terms in the mean are equal to each
other, which happens only if a = d. Similarly, we find

c2011 + · · · + c2011−i bi + · · · b2011

q
2012 2011×2012 2011
> (cb) 2 = (cb) 2 .
2012
Multiplying both inequalities, we obtain
2011 2011
(ad) 2 (cb) 2 <1

which is equivalent to abcd < 1, a contradiction.

3
Problem 3. In triangle ABC the midpoint of BC is called M . Let P be a variable interior
point of the triangle such that ∠CP M = ∠P AB. Let Γ be the circumcircle of triangle ABP .
The line M P intersects Γ a second time in Q. Define R as the reflection of P in the tangent
to Γ in B. Prove that the length |QR| is independent of the position of P inside the triangle.

Solution. We claim |QR| = |BC|, which will clearly imply that quantity |QR| is independent
from the position of P inside triangle 4ABC (and independent from the position of A).
This equality will follow from the equality between triangles 4BP C and 4RBQ. This in turn
will be shown by means of three equalities (two sides and an angle): |BP | = |RB|, |P C| = |BQ|
and ∠BP C = ∠RBQ.
Γ Q

P
α

M C
B

(a) |BP | = |RB|

Obvious since R is the reflection of P in a line going through B.

(b) |P C| = |BQ|
Let U be the fourth vertex of parallelogram BP CU . Then U is on line P Q and ∠BU P =
∠U P C = α. If Q is on the same arc P B as A, then ∠BQP = α, and 4QP U is isosceles;
hence, |BQ| = |BU | = |P C|. On the other way, if Q is on the other arc P B, then
∠BQP and α are supplementary, hence BQU = α, and again 4QP U is isosceles; the
same conclusion follows.

(c) ∠BP C = ∠RBQ

Define T to be the midpoint of P R. Then line BT , tangent to circle Γ in B, splits ∠RBQ
into two parts, ∠RBT and ∠T BQ.
We first show that ∠RBT = α. Indeed, by symmetry, ∠RBT = ∠P BT and, since BT
is tangent to Γ, we have that ∠P BT = ∠P AB (because they both intercept the same
arc PdB on circle Γ), from which our claim follows.

4
 We then show that ∠T BQ = ∠BP M . Indeed, since ∠T BQ and ∠BP Q intercept oppo-
site arcs on circle Γ, they are supplementary and we have ∠T BQ = π − ∠BP Q = ∠BP M .
We finally conclude that

∠RBQ = ∠RBT + ∠T BQ = α + ∠BP M = ∠M P C + ∠BP M = ∠BP C.

We have thus shown 4BP C = 4RBQ, which completes the proof.

Note. Notice that point A does not play any role in the problem except fixing circle Γ (and,
for that reason, the result is also valid when P is chosen outside of triangle 4ABC).

Alternative 1 for (b). The law of sines in triangle 4BQM gives

|BM | |BQ|
= . (2)
sin ∠BQM sin ∠BM Q
Since Q belongs to circle Γ, we have either ∠BQP = ∠BAP = α, hence ∠BQM = ∠M P C,
or these angles are supplementary; in both cases they have equal sines. We also have that
∠BM Q and ∠CM P are supplementary, hence have equal sines. Using these facts along with
|BM | = |M C| transforms (2) into

|M C| |BQ|
=
sin ∠M P C sin ∠CM P
from which the law of sines in triangle 4CP M implies that |BQ| = |P C|.

Alternative 2 for (b). Let S be the second intersection of line CP with circle Γ. Then,
∠BSP = α, so BS and M P are parallel; since M is the midpoint of segment BC, P is the
midpoint of SC. If Q is on the same arc P B as A, then the quadrilateral QP BS is an isosceles
trapezoid, and |QB| = |SP | = |P C|. If Q is on the other arc P B, then the quadrilateral
P QBS is an isosceles trapezoid, and again |QB| = |SP | = |P C|.

5
Problem 4. Yesterday, n ≥ 4 people sat around a round table. Each participant remembers
only who his two neighbours were, but not which one sat on his left and which one sat on
his right. Today, you would like the same people to sit around the same round table so that
each participant has the same two neighbours as yesterday (it is possible that yesterday’s left-
hand side neighbour is today’s right-hand side neighbour). You are allowed to query some of
the participants: if anyone is asked, he will answer by pointing at his two neighbours from
yesterday.

(a) Determine the minimal number f (n) of participants you have to query in order to be
certain to succeed, if later questions must not depend on the outcome of the previous
questions. That is, you have to choose in advance the list of people you are going to
query, before effectively asking any question.

(b) Determine the minimal number g(n) of participants you have to query in order to be
certain to succeed, if later questions may depend on the outcome of previous questions.
That is, you can wait until you get the first answer to choose whom to ask the second
question, and so on.

Solution.

(a) f (n) = n − 3.

• Asking n − 4 questions is not enough since the n − 4 people queried might be sitting
in a consecutive string, in which case the n − 4 answers allow one to sit n − 2 people
in the same positions as yesterday, but there is still an ambiguity among the two
remaining ones.
• Let us show that n − 3 questions suffice. Among the 3 people who are not queried,
at least 2 must sit next to people who have been queried. If exactly 2 do, then both
these people must be neighbours of the third, so that the neighbours of everybody
are known and we are done. If all 3 unqueried people sit next to a queried person,
then at least one of them has two queried neighbours, and again it follows that the
neighbours of everybody are known, so that we are done.

(b) g(n) = n − 1 − n3 ( = n − 1 − n+2

     2n   2n−5 
3
= 3
− 1 = 3
).
Say there is a link between two people if and only if they are neighbours. There are in
total n links, which we all need to identify. By asking a person for his neighbours, we can
discover at most two new links. More precisely, if at any point we query a participant
who has not yet been pointed as a neighbour, we discover exactly two new links (we call
this a type-0 query). If we query a participant who has been pointed once as a neighbour,
will discover exactly one new link (we call this a type-1 query). Of course, querying a
participant who has already been pointed twice provides no information (and we assume
in the rest of this solution that it never happens).
First note that, since f (4) = 1, we also have g(4) = 1. We now prove the formula for
g(n) for n ≥ 5.

6
 n
• Let us show that n −
n 1 − 3
questions suffice. Our strategy consists in making
sure that the first 3 queries are type-0. Let us show that this is always possible.
A type-0 query requires a participant that hasn’t been queried or pointed before.
Since the number of those participants decreases n by three at most after each query,
 n  possible to perform 3 type-0 queries first. During this
we see that it is always
phase we discover 2 3 links.
The remaining queries will be either type-0 or type-1, and each of them discovers at
least one new link. We perform them until n − 1 links have been discovered, after
which we are done (the last link can be deduced without query).n The number of
1
queries
 in
 this second phase  most n − 1 − 2 3 , and the total is at
  is therefore at
most n3 + (n − 1 − 2 n3 ) = n − 1 − n3 .
• We now show that n − 2 − n3 = ĝ(n) questions are not enough.
 

(i) Consider the pool of unqueried and unpointed participants ; each type-0 must
query this pool. Since, from the point of view of the questioner, all elements
of the pool are undistinguishable, we can assume that each type-0 query asks
the second leftmost participant in the pool (except if there is only one element
left in the pool). One can then check that the pool, which starts as a string
of n contiguous participants, will stay contiguous after each type-0 and type-
1 query. Furthermore, using our assumption, we see that each type-0 query
removes three participants from the pool. Therefore there can be at most n3
 

type-0 queries in the scenarios corresponding to our assumption.

(ii) Assume there are k type-0 queries. Since there are ĝ(n) queries, thenumber of
n

discovered links is equal to 2k + (ĝ(n) − k) = ĝ(n) + k = n − 2 + k − 3 . If k is
strictly less than n3 , we discover strictly less than n − 2 links, which is clearly


insufficient (indeed, there are at least three missing links, and one can check
that whatever the configuration of the missing links, there are always several
orders compatible with the discovered links).
(iii) We now analyze the remaining case with k = n3 type-0 queries2 , in which we
 

discover n − 2 links. On the one hand, if the missing links are disjoint, there
are always two orders compatible with the discovered links (for example when
n = 7 and links are missing between the (4, 5) and (7, 1) pairs of neighbours,
the two orders are 1 − 2 − 3 − 4 5 − 6 − 7 and 1 − 2 − 3 − 4 7 − 6 − 5). On the
other hand, a situation where the two missing links would be adjacent would
allow the identification of the correct order. However, this never happens in the
scenarios corresponding to the assumption we made in (i). Indeed, two adjacent
missing links imply that some participant  n is
 unqueried and unpointed at the end
of the process. Since we perform k = 3 type-0 queries (the maximum), the
reasoning from (i) shows that the pool of unqueried and unpointed participants
is empty at the end of the process, which contradicts the existence of two
adjacent missing links.

Here we use the assumption n ≥ 5, since quantity n − 1 − 2 n3  isnegative when n = 4.

1
 
2
Note that this cannot happen when n ∈ {4, 5, 7} since we have n3 > ĝ(n) in those cases.