IrMO_2_2005

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

EIGHTEENTH IRISH MATHEMATICAL OLYMPIAD

Saturday, 7 May 2005


2 p.m.–5 p.m.
Second Paper

1. Let ABC be a triangle, and let X be a point on the side AB that is not A or
B. Let P be the incentre of the triangle ACX, Q the incentre of the triangle
BCX and M the midpoint of the segment P Q. Show that |M C| > |M X|.

2. Using only the digits 1, 2, 3, 4 and 5, two players A, B compose a 2005-digit


number N by selecting one digit at a time as follows: A selects the first digit,
B the second, A the third and so on, in that order. The last to play wins if
and only if N is divisible by 9. Who will win if both players play as well as
possible?

3. Suppose
w
that x is an integer and y, z, w are odd integers. Show that 17 divides
yz yz
x −x .
[Note: Given a sequence of integers an , n = 1, 2, . . ., the terms bn , n = 1, 2, . . .,
a
a a 1
a 1 a 2
of its sequence of “towers”, a1 , aa21 , a32 , a43 , . . ., are defined recursively as
follows: b1 = a1 , bn+1 = abn+1
n
, n = 1, 2, . . ..]

4. Find the first digit to the left, and √


the first
√ digit to the right, of the decimal
2000
point in the decimal expansion of ( 2 + 5) .

5. Let m, n be odd integers such that m2 − n2 + 1 divides n2 − 1. Prove that


m2 − n2 + 1 is a perfect square.
OUTLINE SOLUTIONS

1. Let ABC be a triangle, and let X be a point on the side AB that is not A or
B. Let P be the incentre of the triangle ACX, Q the incentre of the triangle
BCX and M the midpoint of the segment P Q. Show that |M C| > |M X|.
[JC]

Solution. Observe that ∠AXP = 12 ∠AXC and that ∠BXQ = 12 ∠BXC.


Therefore, ∠P XQ = π/2—thus X lies on the circle with diameter P Q. Simi-
larly, ∠P CQ = 12 ∠ACB < π/2—thus C lies outside the circle with diameter
P Q.
2. Using only the digits 1, 2, 3, 4 and 5, two players A, B compose a 2005-digit
number N by selecting one digit at a time as follows: A selects the first digit,
B the second, A the third and so on, in that order. The last to play wins if
and only if N is divisible by 9. Who will win if both players play as well as
possible? [SB]

Solution. Let Nj be the number that has been constructed after j steps, and
let dj be the jth digit, P
so that Nj = d1 d2 d3 . . . dj . A wins if N2005 is divisible
by 9, or equivalently if 2005
j=1 dj ≡ 0 mod 9 .
(1)

Let Wj be the set of remainders mod 9 that guarantees a win for the player
who has just picked the jth digit if that player pursues an optimal strategy
from there on. Thus W2005 = {0}, and we can then work out Wj by backwards
induction using the following rule:

Wj−1 is the set of all remainders r such that r + 1, . . . , r + 5 all fail


to be equivalent to any of the remainders in Wj .

Thus W2004 = {0, 1, 2, 3} and W2003 = {3}(2) .


Notice that the single element in W2003 is 3 larger than the single element in
W2005 . Since the backwards induction rule involves only additions, and always
the same ones, this shift propagates backwards(3) . Now A picks 1003 digits,
1003 ≡ 4 mod 9, and (4 − 1)(3) = 9 ≡ 0 mod 9. Thus W1 = W2005 = {0}.
However N1 = d1 cannot be a multiple of 9, so A will lose if B plays as well
as possible.
Note that B chooses dj for all even j. The winning strategy of B consists of
ensuring that when j is even, Nj is equivalent to one of the remainders in Wj
given as follows:

W2 = W8 = W14 = . . . = W2000 = {6, 7, 8, 0}


W4 = W10 = W16 = . . . = W2002 = {3, 4, 5, 6}
W6 = W12 = W18 = . . . = W2004 = {0, 1, 2, 3}

3. Suppose
w
that x is an integer and y, z, w are odd integers. Show that 17 divides
yz yz
x − x . [KH]

w
Solution. z, w odd ⇒ z w ≡ z (mod 4) ⇒ y z ≡ y z (mod 16) (since y 4 ≡ 1
zw z
(mod 16)) ⇒ xy ≡ xy (mod 17) since either 17|x or x16 ≡ 1 (mod 17).

4. Find the first digit to the left, and √


the first
√ digit to the right, of the decimal
2000
point in the decimal expansion of ( 2 + 5) . [RW]
√ √ √ 1000 √ 1000
Solution. Let x = ( √2 + 5)2000 = √ (7 + 2 10) . Let y = (7 − 2 10) .
Then x + y = m + n 10 + m − n 10 = 2m for some integers m and n. By
the Binomial Theorem,
    
1000 1000 998 2 1000 996 4
2m = 2 7 + 7 · 2 · 10 + 7 · 2 · 102 + · · ·
2 4

+21000 · 10500
≡ 2 · 71000 (mod 10).

Since 

1 (mod 10) if k ≡0 (mod 4)

7 (mod 10) if k ≡1 (mod 4)
7k ≡

9 (mod 10) if k ≡2 (mod 4)


3 (mod 10) if k ≡3 (mod 4)
and 1000 ≡ 0 (mod 4), √ 2m ≡ 2 (mod 10). So the last digit of the integer
x + y is 2. Now 7 − 2 10 is approximately 0.7, so y < 0.01. It follows that
x = (x + y) − y = (. . . 2) − y = . . . 1.9 . . .. The first digit to the left of the
decimal point is 1 and the first digit to the right of the decimal point is 9.

Problem 1 Let n ∈ N. Determine as a function of √ n the


√ first digit to the left
2n
of the decimal point in the decimal expansion of ( 2 + 5) .


Problem 2 Let n ∈ N. Determine the integer part of (2 + 3)n .

5. Let m, n be odd integers such that m2 − n2 + 1 divides n2 − 1. Prove that


m2 − n2 + 1 is a perfect square. [TL]

Solution. We have n2 − 1 = q(m2 − n2 + 1) for some integer q. This can be


written
m2 = (q + 1)(m2 − n2 + 1).

Let q + 1 = k. Observe that it suffices to show that k is a perfect square.


Let S be the set of all ordered pairs of integers (x, y) for which

(x + y)2 = k(1 + 4xy).

Note that the ordered pair of integers ((m + n)/2, (m − n)/2) belongs to S, so
S is not empty. Note also that (x, y) ∈ S implies that (−x, −y) ∈ S and that
(y, x) ∈ S also. Note further that if (0, z) ∈ S
for some integer z, then k is a perfect square, as required.
Suppose then that no integer z exists with (0, z) ∈ S. Let (a, b) ∈ S be such
that | a | is least possible among all pairs (x, y) ∈ S. Note that we may assume
that a =| a | is positive. The quadratic equation

(x + a)2 = k(1 + 4xa)

has two solutions, one of which, b, is an integer. Hence both solutions, b1 and
b2 , say, are integers. Note that b1 + b2 = 4ka − 2a and that b1 b2 = a2 − k.
Hence

(a + b1 )(a + b2 ) = a2 + a(4ka − 2a) + a2 − k = k(4a2 − 1). (1)

Suppose first that k < 0. Then b1 + b2 < 0 and b1 b2 > 0. So b1 and b2 are
both negative. The minimality of a implies that a + b1 and a + b2 are both
nonnegative, so, by (1), k(4a2 − 1) ≥ 0, thus forcing k = 0.
Hence k ≥ 0. If k > 0, then b1 + b2 > 0 and (a + b1 )(a + b2 ) > 0, so, by
minimality of a, b1 and b2 are both positive. But then the fact that b1 b2 =
a2 − k < a2 contradicts the minimality of a. Hence k = 0, and (0, 0) ∈ S. This
is a contradiction. So the result is proved.
Remark. It is difficult to find pairs (m, n) satisfying the hypotheses with
m 6= ±n or n 6= ±1. However, such pairs do exist, for example (m, n) =
(105, 99).

Now try the following problem which was set for the IMO in 1988.

Problem 3 Let a and b be positive integers such that ab + 1 divides a2 + b2 .


Show that
a2 + b2
ab + 1
is the square of an integer.

You might also like