Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

The Blakers Mathematics Competition 2006

PROBLEMS AND SOLUTIONS

1. Corner Reflector

The last American manned lunar expedition installed on


the Moon’s surface an optical device consisting of three
planar mirrors meeting as portions of the internal corner
of a cube. The purpose of this device was to obtain very
accurate measures of the Earth–Moon distance by timing
the transmission and return of a pulse of light. Show that
if a ray of light reflects off all three mirrors, it leaves the
corner in the same direction it came from; i.e., parallel to
the incoming ray.

Solution: Take a coordinate system in which the xy, yz


and zx planes are the three mirrors. An incoming ray is
determined by a unit vector, whose entries are called the
direction cosines of the direction. Assume first that the
incoming ray first meets the xy–plane and has direction
the unit vector (a, b, c).
The laws of optics say that the first reflection, with direc-
tion (a0 , b0 , c0 ) is coplanar with the incoming ray and the
normal to the xy plane at the point of reflection. Hence a0 =
a and b0 = b. Also, the angle of incidence equals the angle
of reflection, so c0 = −c. Hence (a0 , b0 , c0 ) = (a, b, −c).
The same thing happens with the other two reflections;
that is, the direction component normal to the plane is
reversed and the other two are unchanged. Thus if the next
reflection plane is the xz–plane, after the second reflection
the direction cosines are (a, −b, −c) and after the third
reflection they are (−a, −b, −c).
If the second reflection plane is the yz-plane they are
(−a, b, −c) after the second reflection and (−a, −b, −c)
after the third. In any case, the ray emerging after three
reflections is parallel to the incoming ray.
A similar argument holds whatever the order in which the
mirrors are hit.

2. Maximising the side differences in a triangle

Let a, b and c be the side lengths of a triangle with fixed


perimeter 2s. As a, b and c vary, what is

sup{(b − c)2 + (c − a)2 + (a − b)2 }?

Solution: Suppose a ≤ b ≤ c. Consider what happens


when the perimeter and middle side length are held con-
stant while the long side approaches s. The new degenerate
triangle has the same perimeter and middle side and side
lengths s − b, b, s. However, each difference between side
lengths increases.
Thus we have to maximise the poynomial in b, (2b − s)2 +
b2 + (s − b)2 = 6b2 − 6sb + 2s2 , 0 ≤ b ≤ s.
The derivative is 12b − 3s which is zero for b = s/4. So
the critical points are b = 0, s/4, and s and the maximum
value of the polynomial is 2s2 , attained at the endpoints.
Alternative: Maximising a function of two variables.
Firstly, 0 < a < s, otherwise the triangle inequality is
violated. Replacing the inequality with equality yields a
degenerate triangle.
Since c = 2s − a − b, c < s implies a + b > s.
Thus we must maximise f (a, b) = (a − b)2 + (a + 2b − 2s)2 +
(2a+b−2s)2 , defined on the region a < s, b < s, a+b > s.
Equating partial derivatives to zero yields a = b = c =
2s/3, which yields the minimum from the equilateral tri-
angle. So the supremum occurs on the boundary. On
a = s, f (s, b) = (s − b)2 + (2b − s)2 + b2 , with critical
points at b = 0, s/2, s, and f has supremum 2s2 at the
vertices (s, 0), (0, s), (s, s), each of which is a degenerate
case.

3. Balls Taken From A Bag

You have a number of black and white balls in a bag, with a


number of black balls left over. You remove two balls from
the bag and then put one ball back in the bag according to
the following rule:
(a) If the two balls removed are both black, put one black
ball back in the bag.
(b) If the two balls removed are both white, put one black
ball back in the bag.
(c) If one ball removed is black and the other is white, put
one white ball back in the bag.
Repeat this operation. Note that at each stage the number
of balls in the bag is reduced by one, so that eventually
there will be only one ball left.
What is the colour of this ball?

Solution: Notice that if the two balls removed are both


black, or if one is white and the other black, then the num-
ber of white balls left in the bag is unchanged, while if the
two balls removed are both white, then the number of white
balls left in the bag decreases by two. Thus at any stage
the number of white balls left in the bag is unchanged, or
decreases by two. If initially there was an odd number of
white balls there will always be an odd number of white
balls, so that when there are just two balls left, one must
be white, the other black. When these are withdrawn, the
one replaced is white. If initially there was an even number
of white balls there will always be an even number of white
balls, so that when there are just two balls left, both must
be white, or both black. When these are withdrawn, the
one replaced is black. Thus if initially there was an odd
number of white balls in the bag, the last one is white; if
initially there was an even number of white balls in the bag,
the last one is black.

4. Coin on a Chessboard
The diameter d of a coin is less than the side length s of a
single square on an infinite chessboard.
What is the probability that when the coin is rolled on the
chessboard and comes to rest on the board that it covers
some corner of one of the squares?
Solution The centre of the coin lies within a square or on
the boundary of a unique side. Since d < s, the coin can
cover at most one corner of a square. In order to do so,
its centre must lie within a quadrant of radius d/2 of that
corner. hence it must lie within one of four disjoint regions
of total area π d2 /4 within a square of area s2 .
Therefore probability that the coin covers a corner is πd2 /4s2 .

5. Integral Points on a Curve


How many points on the curve whose equation is
2x2 − 5xy + 4x + 3y = 2
have integer coordinates?

Solution: Suppose (x, y) is such a point. Since


2x2 + 4x − 2 10x2 + 20x − 10 26x − 10
y= , 5y = = 2x +
5x − 3 5x − 3 5x − 2
Hence
130x − 50 28
25y = 10x + = 10x + 26 +
5x − 3 5x − 3
Since y is an integer,(5x−3)|28 so 5x−3 = ±1, ±2, ±4, ±7, ±14, ±28.
Since x is an integer, the only possibilities are x = 1, 2 or
−5.
If x = 1 then y = 2. If x = 2 then y = 2. if x = −5 then
y = −1. Hence there are three integer points on the curve,
(x, y) = (1, 2), (2, 2) and (−5, −1).
6. Peculiar Polynomials

Call a monic polynomial peculiar if its coefficients are in


arithmetic progression and its roots are integers. One ex-
ample is x2 − 1, whose coefficients are 1, 0, -1 with roots -1
and 1. Find all peculiar polynomials of degree 2.

Solution: (Hiro Asihara, Y1, UWA)


A monic quadratic with its coefficients in arithmetic pro-
gression has formula x2 + (k + 1)x + 2k + 1 for some in-
teger k. For the roots to be integers, the discriminant
k 2 − 6k − 3 = ∆2 for some integer ∆.
√ q
Hence k = 3 ± ∆ + 12 = 3 ± ∆ 1 + ∆122 . This implies
2

that ∆2 + 12 is a square and ∆2 divides 12.


The only possibility is ∆ = 2 so k = −1 or 7.
Hence the only peculiar polynomials of degree 2 are x2 − 1
and x2 + 8x + 15.
Alternative Evgeni Sergeev, Y3, UWA pointed out that
there are other solutions, since we did not state in which
order the coefficients had to be in arithmetic progression.
For example, in addition to the given solutions, there are
the quadratics x2 + 2x, x2 − 6x + 8, x2 + 3x + 2 and x2 − x.

7. A Matrix Operation

Let A, B and C be n × n matrices over the reals satisfying


A + B = AB, B + C = BC and C + A = CA. What is
(A − In )2 + (B − In )2 + (C − In )2 ? Find non–zero examples
of such matrices.

Solution: (Evgeni Sergeev, Y3, UWA) From A + B = AB


and B + C = BC we have AC + BC = ABC = AB + AC.
Hence AB = BC and similarly, BC = CA. Hence A + B =
B + C = C + A so A = B = C.
Hence A2 = 2A, B 2 = 2B and C 2 = 2C, so (A − In )2 +
(B − In )2 + (C − In )2 = 3In .
From A2 = 2A, the diagonal matrices whose diagonal ele-
ments are all either 0 or 2 provides 2n −1 non-zero examples
for all positive integers n.
For other examples, let Q be any invertible n×n matrix and
E any given example. Then QEQ−1 is another example.

8. Lottery Results

On 29 June 1995 the German press reported the following


sensation: “For the first time it its history a lottery drawing
resulted in winning numbers that had been already drawn
once. The numbers drawn on 21 June 1995, namely 15–25–
27–30–42–48, were identical to those drawn on 20 December
1986. There are nearly 14 million different combinations
that can be drawn.”
To put this into context, the German lottery player draws
six numbers (without replacement) out of 49. And the
drawing on 21 June 1995 was the 3016th draw of the lottery.
Is this fact really sensational? What is the probability that
the first repeat does not happen until the 3016th draw?
Solution: (Wilson Ong, Y1, UWA)

P (First repeat does not occur before 3016-th draw)


=P (No repeats in first 3015 draws)
49
 49 49
 49

6 − 1 − 2 − 3014
= 49  · 6 49 · 6 49 · · · · . 6 49
6 6 ! 6 ! 6 !
1 2 3014
=1. 1 − 49 . 1 − 49 . . . . . 1 − 49
6 6 6
1
49 x

x
−1 −2 −3014
≈(e .e ···e ) 6 since 1 + ≈ e n for large n.
n
4543605
=e− 13983816 ≈ 0.7226
Hence P (First repeat occurs within first 3015-th draws)
≈1 − 0.7226 = 0.2774

Therefore the fact that the first repeat occurred on the


3016th draw is not really sensational.

9. A Wind Chime

A wind chime consists of a circular disc, suspended on


a cord attached to the centre, and weights 1, 2, 3, . . . , 10
hung at equal spacing around the circumference. If the
weights were hung in the order 1, 2, 3, . . . , 10, then of course
the disc would not be horizontal. Is it possible to hang the
weights in such an order that the disc is horizontal? If so,
list the order of the weights: if not, give a proof why not.

[It may be helpful to know that cos 36◦ = ( 5 + 1)/4]
Solution: (Michael Bausor, Y2, UWA)
Consider two such discs, each with weights distributed as
shown in the diagrams below.

1 5
5 2 0 0

4 3 5 5

3 4 0 0

2 5 5 5

1 0

By symmetry, each of the dics suspended on a cord attached


to the centre remains balanced. Hence when they are placed
on top of each other, with a 5 on top of a 5 they will still be
balanced with the weights in order 1, 7, 3, 9, 5, 6, 2, 8, 4, 10.

10. Sum of a Series

Find the sum and the region of convergence of the series:


2x2 3x3 4x4 (−1)n+1 (n + 1)xn+1
− + + ··· + + ···
1.3 2.4 3.5 n(n + 2)

Solution: (Wilson Ong, Y1, UWA)


Let
(−1)n+1 (n + 1)xn+1 (−x)n+1 (n + 1)
an = =
n(n + 2) n(n + 2)
n n
 
(−x) (−x)
=−x +
n+2 n(n + 2)
n
(−x)n
 
x (−x)
=− +
2 n+2 n
P∞
If x = 0 then clearly n=1 an = 0, so assume x 6= 0.
Then
1 (−x)n+2 (−x)n (−x)2
 
an = − + so
2x n+2 n
∞ ∞ ∞
X 1 X (−x)n+2 x X (−x)n
an = − −
n=1
2x n=1
n + 2 2 n=1 n

Hence the region of convergence of the series is (−1, 1].


(Note: This could also be verified by using the ratio test to
find the interval of convergence, the alternating series test
for x = 1 and the harmonic series for x = −1.)
For x ∈ (−1, 0) ∪ (0, 1], since
x2 x3 (−1)n+1 xn
x− + + ··· + + · · · = ln(x + 1)
2 3 n


x2
 
X 1 x
an = ln(x + 1) − x + + ln(x + 1)
n=1
2x 2 2
 
1 x x 1
= + ln(x + 1) + −
2x 2 4 2
Hence

(
X 0 for x = 0
an = 1 x
 x 1
n=1 2x + 2 ln(x + 1) + 4 − 2 for x ∈ (−1, 0) ∪ (0, 1]

You might also like