Blakers 2009

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

The Blakers Mathematics Contest 2009 Solutions

1. Interesting function
1
Let y = x x , x > 0. Find the intervals on which y(x) is
monotonic, and on each such interval, find its range.

Solution: [Robert Palmer, Y2, UWA] Since x > 0 implies


y > 0, we can take logs on both sides to conclude
1
ln y = ln x
x
Differentiating with respect to x, we get
1 dy 1
= 2 (1 − ln x)
y dx x
Since y and x2 > 0, the only crtical point is at ln x = 1, x = e.
dy
Furthermore, dx is positive for 0 < x < e and negative for
e < x < ∞.
Hence y(x) is increasing on (0, e] with range (0, e1/e ] and de-
creasing on [e, ∞) with range [e1/e , 1).

2. Approximate area
A segment of a circle is the smaller region cut off the circle
by a chord. The apothem of the segment is the line segment
perpendicular to the chord from the midpoint of the chord to
the perimeter of the segment. Two methods for approximating
the area of the segment have been proposed:
(a) The trapezoidal approximation is the area of the trapezoid
whose base is the chord, and upper edge the tangent to
the segment at its midpoint with length the length of the
apothem.
(b) The parabolic approximation is the area of the segment of
the parabola passing through the endpoints of the chord
and the vertex of the segment.
Show that the trapezoidal approximation is more accurate than
the parabolic when the segment is ‘large’ (approaches a semicir-
cle), while the parabolic approximation is more accurate when
the segment is ‘small’.
Find the segment for which the trapezoidal and parabolic
approximations are equal.

1
Solution: Without loss of generality, assume the radius is
1. Let c be the length of the chord and h the length of the
apothem and θ half the angle subtended by the chord at the
centre of the circle.
Then the trapezoidal approximation is A(t) = 21 (c + h)h and
the parabolic approximation is A(p) = 23 ch.
These are equal when h = 31 c, and in this case, A(t) = A(p) =
2 2
9
c.
The actual area A = θ − sin θ cos θ.
As θ → π2 , c → 2 and h → 1 so A(t) → 32 , A(p) → 43 while
A → π2 ≈ 1.57, so the trapezoidal approximation wins for large
segments.
For the limit as θ → 0+ we need to express A(t) and A(p) in
terms of θ. Since c = 2 sin θ and h = 1 − cos θ,

1
A(t) = (2 sin θ + 1 − cos θ)(1 − cos θ)
2
1
= (2(sin θ − cos θ) − sin 2θ + cos2 θ + 1) and
2
2 2
A(p) = (2 sin θ)(1 − cos θ) = (2 sin θ − sin 2θ).
3 3

Since each of limθ→0+ A, limθ→0+ A(t) and limθ→0+ A(p) = 0,


we must use L’Hospital’s Rule to compute limθ→0+ A/A(t) and
limθ→0+ A/A(p).
To simplify the notation, we first compute derivatives of each
of the three functions until the limit is non–zero.

1
A=θ− sin 2θ → 0, A0 = 1 − cos 2θ → 0, A00 = 2 sin 2θ → 0,
2
A000 = 4 cos 2θ → 4.
1
A(t) = (2 sin θ − sin 2θ − 2 cos θ + cos2 θ + 1) → 0,
2
1
A(t)0 = cos θ + sin θ − cos 2θ − sin 2θ → 0,
2
00
A(t) = − sin θ + cos θ + 2 sin 2θ − cos 2θ → 0,
A(t)000 = − cos θ − sin θ + 4 cos 2θ + 2 sin 2θ → 3.
2 2
A(p) = (2 sin θ − sin 2θ) → 0, A(p)0 = (2 cos θ − 2 cos 2θ) → 0,
3 3
2 2
A(p)00 = (−2 sin θ + 4 sin 2θ) → 0, A(p)000 = (−2 cos θ + 8 cos 2θ) → 4.
3 3
So miraculously, all three third derivatives have non–zero lim-
its so by L’Hospital’s Theorem, limθ→0+ A/A(t) = 4/3 while
limθ→0+ A/A(p) = 1.
Hence the parabolic approximation wins for small segments.

3. Matrix Inverse
Let A and B be n × n real matrices satisfying
(a) AB = BA
(b) A2 = A and B 2 = B
(c) A − B is invertible.
Prove that A + B = In , the identity matrix.

Solution: [Aaran Mohann, Y2, Curtin]


(A + B)(A − B) = A2 − AB + BA − B 2
= A2 − B 2 (since AB = BA)
= A − B (since A = A2 and B = B 2 ).
Since A − B is invertible,
A + B = (A + B)(A − B)(A − B)−1
= (A − B)(A − B)−1 = In .

4. Integer divisors
(a) Prove that if x and y are positive integers, then xy divides
x2 + y 2 if and only if x = y.
(b) Prove that there are infinitely many triples (x, y, z) of pos-
itive integers such that xyz = x2 + y 2 + z 2 .

Solution: (a) xy divides x2 + y 2 if and only if there is a


positive integer a such that x2 + y 2 = axy.
Consider this as a quadratic in x. It has an integer solution
if and only if the discriminant a2 y 2 − 4y 2 is a perfect square,
which is true if and only if a2 − 4 is a perfect square. Since a is
positive, this holds if and only if a = 2, which is the case if and
only if x = y.
(b) (x, y, z) = (3, 3, 3) is clearly a solution, and if (a, b, c) is
any solution with a ≤ b ≤ c, then so is (b, c, bc − a) because
b2 + c2 + (bc − a)2 = b2 c2 − abc = bc(bc − a). Since bc − a >
c, this leads to an infinite sequence of solutions, for example
(3, 3, 3), (3, 3, 6), (3, 6, 15). . . . .

5. The camel inheritance


An old man willed that upon his death, his three sons would
receive the uth, vth and wth parts of his herd of camels (i.e.,
1/u, 1/v and 1/w of the herd). He had uvw − 1 camels in his
herd when he died. Since they could not divide uvw−1 into u, v
and w parts, they approached a distinguished mathematician
for help. He rode over and added his camel to the herd, and
then fulfilled the old man’s wishes. One camel remained, which
was of course his own.
How many camels were there in the herd?

Solution: The answer is 41. We may assume that u ≤ v ≤


w, and of course u ≥ 2, otherwise one son would get all the
camels. Then the condition of the will is that vw + uw + uv =
uvw − 1.
(a) If u = 2 then the equation becomes (v − 2)(w − 2) = 5
which has the unique integral solution v = 3, w = 7 and
uvw − 1 = 41.
(b) If u = 3, the equation becomes (2v − 3)(2w − 3) = 11,
which has unique solution v = 2, w = 7. But his violates
the condition u ≤ v.
(c) If u ≥ 4 then we have 1−1/(uvw) = 1/u+1/v+1/w ≤ 3/4,
so uvw ≤ 4 which is impossible.

6. A functional equation
let f : R2 → R be a function satisfying
f (x, y) + f (y, z) + f (z, x) = 0 for all real numbers x, y and z
Prove that there exists a function g : R → R such that f (x, y) =
g(x) − g(y) for all real numbers x and y.

Solution: [Robert Palmer, Y2, UWA] From (x, y, z) = (0, 0, 0),


we obtain 3f (0, 0) = 0 and hence f (0, 0) = 0.
From (x, y, z) = (0, a, 0), we obtain for all a ∈ R, f (0, a) +
f (0, 0) + f (a, 0) = f (a, 0) + f (0, a) = 0, or f (0, a) = −f (a, 0).
Hence for all x, y ∈ R, f (x, y) + f (y, 0) − f (x, 0) = 0.
Define g(x) = f (x, 0). Then for all x, y ∈ R, f (x, y) =
g(x) − g(y).

7. Average speed
A racing motorist completed n laps of a course at average lap
speeds of v1 ≤ v2 ≤ · · · ≤ vn . Show that the average speed for
the n laps satisfies v1 ≤ V < nv1 , but there is no lower bound
on V as a function of vn .

Solution: Clearly v1 ≤ V . Suppose the length of the course


is d. Then
total distance nd nv1
V = = Pn 1 = Pn v1
total time d j=1 vj j=1 vj
Pn v1 nv1
But j=1 > 1 and V = Pn v1 < nv1 .
vj j=1 vj
There is no such lower bound, because for any speeds on the
first n − 1 laps, the last lap can be made sufficiently slowly to
make V as small as you like.

8. An Improper Integral
Show that
Z T
lim sin(x) sin(x2 ) dx
T →∞ 1
converges.

Solution:
Integrating by parts
Z T
sin(x) sin(x2 ) dx
1
− cos x2 − cos x2
   Z   
sin x − sin x cos x
= − + dx
x 2 2 x2 x
− cos x2 cos x2 sin x
   Z
sin x
= − dx
x 2 2x2
cos x sin x2 sin x2
  Z    
1 − sin x 2 cos x
+ − − dx
2x2 2 2 2 x2 x3
Since
sin T cos T 2 cos T sin T 2
lim = 0 = lim
T →∞ T T →∞ T
we only have to show that
Z T Z T Z T
cos x2 sin x sin x2 sin x sin x2 cos x
dx, dx, and dx
1 x2 1 x2 1 x3
all converge as T → ∞. But all three converge absolutely since
RT 1 RT 1
1
dx and 1
dx both converge as T → ∞.
x2 x3
Alternative Solution: [Wilson Ong, Y3, UWA]
2
By Euler’s identity, eix = cos x2 + i sin x2 .
R ∞ ix2 R∞ 2
R∞
0
e dx = 0R
cos x dx + i sin x2 dx converges since the
0 R
∞ ∞
Fresnel integrals 0 cos x2 dx and 0 sin x2 dx both converge.

eix = cos x + i sin x and e−ix = cos x − i sin x imply sin x =


e−ix −eix
2
i.

Now
Z ∞ Z ∞ 
2 2
sin x sin x dx = = sin x(i sin x ) dx
1 1
Z ∞ 
ix2 2
= = sin x(e − cos x ) dx
1
Z ∞ 
ix2
= = e sin x dx
1
Z ∞ 
ix2 e−ix −eix
= = e ( 2 i) dx
1
 Z ∞ 
1 ix2 −ix ix2 +ix
= = i (e −e ) dx
2 1
 Z ∞ Z ∞ 
1 −i/4 i(x−1/2)2 −i/4 i(x+1/2)2
= = i(e e dx − e e dx)
2 1 1
R∞ 2 R∞ 2 R∞ 2
Observe that 1 ei(x−1/2) dx = 1/2 eiu du = 0 eiu du −
R 1/2 iu2 R ∞ iu2 R 1/2 iu2
e du is convergent since e du and e du =
R01/2 2
R 1/2 2
0 0

0
cos u du + i 0 sin u du are convergent.
R∞ 2 R∞ 2 R∞ 2 R 3/2 2
Similarly, 1 ei(x+1/2) dx = 3/2 eiu du = 0 eiu du− 0 eiu du
is convergent.
R∞
Hence 1
sin x sin x2 dx converges.

9. Cakes and Boxes

A triangular cake and a triangular box are congruent, but


mirror images of each other. We would like to cut the cake into
two pieces which can fit together in the box without turning
either piece over.
(a) Show that this is possible if one angle of the triangle is
three times as large as another.
(b) Show that it is possible if one angle of the triangle is obtuse
and twice as large as another.

Solution: [Wilson Ong, Y3, UWA] Denote the triangle by


ABC and let θ be the angle at A.
A A0
@
θ @
D B 0 XX@
B  θ  X @ D0
X
B 2θ 2θ 
B 
B 
B 
B 
B 
B 
B 
B 
B 
B 
B 
C C0
Cake Box
(a) Let the angle at B be 3θ. Let D be the point on AC such
that ∠ABD = θ. Then the triangle ABD is isosceles with
AD = BD. Also, ∠BDC = ∠DBC = 2θ, so triangle
BCD is isosceles with BC = DC.
Cut off triangle ABD and translate and rotate it so that
A moves to D0 and D moves to B 0 .
A A0

 θ
  Q
Q
E Q E0
 θXX
X Q
 Q

2θ
BA 2θ D D0 Q 0
B
A 
A 
A 
A 
A 
A 
A 
A 
A 
A 
C C0
Cake Box
Let the angle at B be 2θ. Let D be the point on AC
such that BC = DC and let E be the point on AB such
that ∠AED = θ. Then the triangle AED is isosceles with
AD = ED. Also, ∠BDE = ∠DBE, so triangle BDE is
isosceles with BE = DE.
Cut off triangle AED and translate and rotate it so that
A moves to E 0 and E moves to A0 .

10. Centre of mass


The centre of mass of a can of Coke (which may be idealised
as a cylinder of height 2h and radius a and flat ends) is h above
the base when the can is full or empty. Suppose the can is
partly full. Show that the centre of mass is lowest when it lies
in the surface of the liquid.
Does this result change when the can is replaced by a bottle
of traditional Coke shape?

Solution: [Adrian Dudek, Y3, UWA] Let C(y) be the height


of the centre of mass when the height of the liquid is y, 0 ≤
y ≤ 2h. Then C(0) = h, C(2h) = h.
When the liquid is removed from a full can, the centre of mass
falls; likewise when liquid is added to an empty can. Since
C(y) is a continuous function, by Rolle’s theorem, there is a
value of y in 0 ≤ y ≤ 2h where C 0 (y) = 0. That is, there is a
minimum value of C(y) which is less than h. Thus C(y) ≤ h for
0 ≤ y ≤ 2h so by the intermediate value theorem there exists
a value y for which C(y) = y; that is, the centre of mass lies in
the surface.
When the centre of mass lies in the surface, if liquid is added,
the height of the centre of mass increases, since adding mass
above the existing centre of mass must raise the centre of mass.
Likewise, when the centre of mass lies in the surface, if liquid
is removed, then the height of the centre of mass increases, since
removing mass below the existing centre of mass must raise the
centre of mass.
Hence the centre of mass is a minimum when it lies in the
surface.
For a Coke bottle shape, the same argument applies although
the height of the centre of mass of a full bottle is different from
that of an empty bottle.
Alternative Solution
Suppose the shape of of the container is given by rotation of
the curve x = f (y) about the y–axis. (For the cylindrical can,
f (y) = a.) Let the mass of the container be M and its empty
centre of mass at height h.
The contribution of the container to the first moment of mass
about the base is M h.
The contribution to the first moment of the layer of fluid of
thickness dy at height y is yρπf 2 (y) dy, where ρ is the density,
and the total contribution of the liquid is the integral of this.
Hence the centre of mass is
Ry Ry 2
0R
uρπf 2 (u) du + M h 0R
uf (u) du + Kh
C(y) = y = y
0
ρπf 2 (u) du + M 0
f 2 (u) du + K
where M = ρπK. R
y R y 2
C 0 (y) = 0 when 0 Rf 2 (u) du + K [yf 2 (y)]
  2
Ry 2 0= uf (u) du + Kh [f (y)].
2 y 2
Since f (y) 6= 0, y 0 f (u) du + Ky = 0 uf (u) du + Kh.
Denote the solution of this equation by y ∗ , where
R y∗ 2
uf (u) du + Kh
y ∗ = 0R y∗ = C(y ∗ )
2
f (u) du + K
0
which says that the height of the surface and the minimum
height of the centre of mass coincide.

You might also like