2012 Bored of Studies Trial Examinations

Mathematics Extension 2
SOLUTIONS

Disclaimer: These solutions may contain small errors. If any are found, please feel free to
contact either Carrotsticks or Trebla on www.boredofstudies.org, regarding them.

Thanks: To Trebla, for his many hours spent verifying solutions and suggesting alternate
methods.
Multiple Choice
1. A
2. A
3. D
4. C
5. C
6. B
7. B
8. D
9. A
10. A

Brief Explanations
Question 1 Definition of ellipse.

Question 2 Vector must be longer, since we have the reciprocal of a vector that has
modulus less than 1. The arguments must also be subtracted, since we are dividing two
vectors.

Question 3 Standard t substitution integral.

Question 4 Recall that  n  1   n  1  0    1  n1   n2  ...    1  0 . So we

n 1
1  2k  1
therefore have   k  0 . Also, a further manipulation  k 
k 0 k

k
along with

 n1   n  1   yields the final result.

Consider P  x    x  k  Q  x  , where k  0 and n  .

n
Question 5

Question 6 Standard resolution of forces. Note that F points upwards since V is less than
the optimal speed.

Question 7 Use the inequality a2  b2  c2  ab  bc  ac combined with

a  b  c  a 2  b2  c 2  2  ab  bc  ac  to acquire the answer. Note that a  b  c  1.
2

Question 8 (B) and (C) are not even. (A) diverges.

Question 9 The locus is just xy  1 .

Question 10 Use the Conjugate Root Theorem to factorise, then the discriminant.

–1–
Written Response
Question 11 (a)

First divide numerator and denominator by e x , allowed since e x  0 for all x  .

1
 e x  e x
1
 e2 x  1
 2x dx   x  x dx
0 e  1 0 e  e
1

 ln  e  e
x x

0

ee  1
 ln  
 11 
 e2  1 
 ln  
 2e 

Question 11 (b) (i)

Cross multiply the expressions.

1 A  x  3cos x  2sin x   B  x  3cos x  2sin x 


9cos x  4sin x
2 2
9cos 2 x  4sin 2 x

But recall that sin 2 x  cos2 x  1 , so we therefore have, by equating the numerator:

sin 2 x  cos2 x  A  x  3cos x  2sin x   B  x  3cos x  2sin x 

Group the sines and cosines:

sin 2 x  cos2 x  2sin x  A  x   B  x    3 cos x  A  x   B  x  

So therefore, we have:

1 1
A  x   B  x   cos x A  x   B  x   sin x
3 2

Solving simultaneously, we immediately acquire:

11 1  1
A  x    cos x  sin x    2 cos x  3 sin x 
23 2  12

Similarly:

11 1  1
B  x    cos x  sin x    2 cos x  3 sin x 
23 2  12

–2–
Question 11 (b) (ii)

Our integral is now:

 
 4 dx 1  4  2 cos x  3 sin x 2 cos x  3 sin x 
    
0 9 cos x  4sin x 12 0  3cos x  2sin x 3cos x  2sin x 
2 2
dx

1  4  3sin x  2 cos x 3sin x  2 cos x 
     dx
12 0  3cos x  2sin x 3cos x  2sin x 

1  4  3sin x  2 cos x 3sin x  2 cos x 
     dx
12 0  3cos x  2sin x 3cos x  2sin x 

1  4
 ln  3cos x  2sin x   ln  3cos x  2sin x  
12  0

1   3cos x  2sin x   4
 ln  
12   3cos x  2sin x   0

  3 2  4

1   2   ln 1 
 ln  2 
12   3  2  
  2 2
  0
1
 ln  5   ln 1 
12
1
 ln 5
12

–3–
Question 11 (c) (i)

R
r

Computing the volume of the shell, we have:

 Vn    R 2  r 2  x  x n 

 1  x    x   1  x    x  x 
2 2 n


  1  x   2 1  x   x   x   1  x 
2 2 2
 x  x 
n

   2 1  x   x   x    x  x 
2 n

 2 1  x   x  x n   x
 2  x  x n  x 2  x n 1   x
 2  x  x 2  x n  x n 1   x

Summing the shells, we have:

1
Vn  lim  2  x  x 2  x n  x n 1   x 1 1
 2  x 2  x3 
1 n 1
x 
1 n2 
x 
 x 0
0 x 1 2 3 n 1 n2 0
1
 1 1 1 
  2  x  x 2  x n  x n 1   x
1
 2    
0  2 3 n  1 n  2 

1
1 1 1 
 2   x  x 2  x n  x n 1   x  2   
0  6 n  1 n  2 

–4–
Question 11 (c) (ii)

Take the limit as n   .

1
lim Vn  2 
n  6


3

Question 11 (c) (iii)

We observe what happens to y  x n as n   .

Since 0  x  1 , we have 0  y  1 . We know that as n gets large, the curve y  x n gets lower,
in the domain 0  x  1 .

Hence we see that if n gets large, then the area between the two curves will approach a right-
angled triangle (the lone curve on the right, in the above diagram, is y  x 200 ).

So if we rotate this ‘triangle’ about the line x  1 , we will acquire a cone with unit base
radius, and unit height.

1 
Using the formula Vcone   r 2 h , where r  1 and h  1 , we indeed acquire Vcone  .
3 3

–5–
Question 11 (d)

–6–
Question 12 (a) (i)

Differentiate the expression implicitly.

x y  c

1 y
 0
2 x 2 y
1 y
 0
x y
y 1

y x
y
y  
x
y

x

–7–
Question 12 (a) (ii)

Using the point-gradient formula:

y1
y  y1    x  x1 
x1

The line passes through the point  a, 0  :

y1
 y1    a  x1 
x1
1
y1   a  x1 
x1
a  x1  x1 y1
a  x1 y1  x1
 x1  x1  y1 
 x1  c

Similarly for the point  0,b  :

y1
b  y1     x1 
x1
 x1 y1
b  x1 y1  y1
 y1  x1  y1 
 y1  c

So adding them, we have:

a  b  x1  c  y1  c
 c  x1  y1 
 c c
c

–8–
Alternatively

y1
y  y1    x  x1 
x1

We know this passes through  a, 0  :

y1
 y1    a  x1 
x1
a  x1
y1 
x1
x1 y1  a  x1 1
Similarly, it passes through  0,b  :

y1
b  y1     x1 
x1
b  y1  x1 y1  2
Add the two equations:

2 x1 y1  a  b  x1  y1
 a  b   x1  y1 
a  b  x1  2 x1 y1  y1

 x 
2
 1 y1

 c
2

c

–9–
Question 12 (b)

Consider the following diagram.

mg

Resolving forces vertically: N cos 2  mg

Resolving forces horizontally: N sin 2  mr 2

r 2 r
Divide the two expressions to acquire tan 2  . But we also know that tan 2  ,
g R  h1
using the right-angled triangle. Equating the two exprssions yields:

r 2 r

g R  h1
2 1

g R  h1
g
2 
R  h1

Our expression is in terms of R. The required expression is in terms of r, so we will use

Pythagoras’ Theorem to form a relatonship between the two variables.

r 2  R 2   R  h1 
2

 R 2  R 2  2 Rh1  h12
 2 Rh1  h12
r 2  h12
R
2h1

– 10 –
 g
Substitute this into our expression  2  :
R  h1

g
2 
r  h12
2
 h1
2h1
g

r  h12  2h12
2

2h1
2 gh1

r  h12
2

So therefore to maintain a height greater than h1 , the particle must have angular velocity
2gh1

r  h12
2

2gh2
Similarly for h2 , the particle must have angular velocity   .
r  h2 2
2

Therefore we have the required expression:

2 gh1 2 gh
  2 2 2
r  h1
2 2
r  h2

– 11 –
Question 12 (c) (i)

We simply extend the real component one unit forwards and backwards.
y

Note that as a consequence z, z  1 and z  1 are now all collinear.

Question 12 (c) (ii)


We will find Arg  z  1 and Arg  z  1 , and show that they differ by .
2
y


All we need to really spot, is that Arg  z   Arg  z  1  , since the triangle enclosed by the
2
origin, z and z  1 , is an isosceles triangle. A basic angle chase using the exterior angle
 
theorem and the angle sum of triangle quickly yields Arg  z  1  Arg  z   , where it
2

follows immediately that Arg  z  1  Arg  z  1  and thus the result.
2

– 12 –
Alternatively
y

We see immediately that since we have a semi-circle, the angle subtended by the points 1 , 1
and z form a right-angled triangle (Thales’ Theorem), so therefore we have


Arg  z  1  Arg  z  1  , and thus the result.
2

Question 12 (c) (iii)

z 1  
The required statement is equivalent to proving that  tan   , noting that Arg  z    .
z 1 2

From diagram, we can easily see that:

   z 1
tan   
 2  z 1
z 1

z 1

– 13 –
Question 12 (d) (i)

1 1
We know that BOC   BC  OA ' . We also know that ABC   BC  AA ' .
2 2

Hence putting the expressions in a fraction, we immediately acquire:

1
BOC  BC  OA '
 2
ABC 1  BC  AA '
2
OA '

AA '

Question 12 (d) (ii)

OA ' BOC
From (i), we have  .
AA ' ABC

OB ' AOC OC ' AOB

Repeat for other sides to obtain  and  .
BB ' ABC CC ' ABC

Add all the expressions to acquire:

OA ' OB ' OC ' BOC AOC AOB

    
AA ' BB ' CC ' ABC ABC ABC
BOC  AOC  AOB

ABC
ABC

ABC
1

– 14 –
Question 13 (a) (i)

Consider the line y  mx  c . Let this line be tangential to the hyperbola.

To do so, we will substitute the line into the equation of the hyperbola, then let the
discriminant be zero

The equation of the hyperbola is b2 x 2  a 2 y 2  a 2b2 . Substitute y  mx  c :

b2 x 2  a 2  mx  c   a 2b2
2

Expand and re-arrange as a quadratic in terms of x:

b 2 x 2  a 2  m2 x 2  2mcx  c 2   a 2b 2
b 2 x 2  a 2 m2 x 2  2a 2 mcx  a 2c 2  a 2b 2
x 2  b 2  a 2 m2   2a 2 mcx  a 2  b 2  c 2   0

Let the discriminant be equal to zero:

4a 4 m 2c 2  4a 2  b 2  a 2 m 2  b 2  c 2   0
a 2 m 2c 2   b 2  a 2 m 2  b 2  c 2   0
a 2 m 2 c 2  b 4  b 2 c 2  a 2 m 2b 2  a 2 m 2 c 2  0
b 4  b 2 c 2  a 2 m 2b 2  0
b2  c 2  a 2 m2  0
c 2  a 2 m2  b2

Substitute back into y  mx  c :

y  mx  a 2 m2  b2

– 15 –
Question 13 (a) (ii)

We will re-arrange y  mx  a 2 m2  b2 be a quadratic in terms of m.

 a 2 m2  b 2  mx  y
a 2 m2  b 2   mx  y 
2

a 2 m2  b 2  m2 x 2  2mxy  y 2

Hence our quadratic is m2  a 2  x 2   2mxy   b2  y 2   0 .

Since the two tangents are to be perpendicular to each other, we will let the product of roots
be equal to 1 , such that m1m2  1 .

So immediately, we acquire:

  b2  y 2 
 1
a2  x2
b2  y 2
1
a2  x2
b2  y 2  a 2  x 2
x2  y 2  a 2  b2

– 16 –
Question 13 (a) (iii)

We know that OT is the radius of the circle, so OT 2  a 2  b2 .

We also know that OS  ae , since S is the focus of the ellipse.

Squaring both sides, we have OS 2  a 2e2 . However, recall the formula b2  a 2 1  e2  :

b2
1  e2 
a2
b2
e  1 2
2

a
a  b2
2

a2

So therefore:

OS 2  a 2 e 2
 a 2  b2 
 a2  2 
 a 
 a b
2 2

 OT 2

Thus OS  OT , so therefore triangle OTS is isosceles.

– 17 –
Question 13 (b) (i)

Balancing the forces, we have:

ma  mg  mkV 4
a    g  kV 4 

   g  kV 4 
dV
V
dx

Re-arrange and integrate both sides, using appropriate limits:

x V
 
t t
V
 dx   dV
0 U g  kV
4

U
 V
xt   dV
Vt g  kV
4

To integrate this, we let y  V 2 :

dy
dy  2V dV   V dV
2

Finding limits:

V U  y U2
V  Vt  y  Vt 2

So our integral becomes:

– 18 –
 U2
1 dy
x 
2 Vt 2 g  ky 2
U2
1 y k
 tan 1 
 g 
2 kg   Vt 2

1  1  U 2 k   2 k
1 Vt

  tan    tan   
2 kg   g   g  
 U 2 k Vt 2 k 
  
1 1  g g 
 tan
2 kg  U k Vt 2 k
2 
 1   
 g g 
 k 
 U 2  Vt 2  
tan 1  
1 g

2 kg  kU 2 2
V 
 1  
t

 g 
 k 
 U 2  Vt 2  
tan 1  
1 g

2 kg  g  kU 2 2
V 
 
t

 g 
 


1
tan 1 
 t 
 k U 2 V 2 

2 kg  g  kU Vt 
2 2

 
 g 
1  kg U  Vt  
2 2

 tan 1  
2 kg  g  kU 2Vt 2 
 

– 19 –
Question 13 (b) (iii)

For maximum height, let V  0 :

1  U 2 kg 
x tan 1  
 g
2 kg  
1  2 k
1 U

 tan  

2 kg  g 
1  k
 tan 1  U 2 
2 kg  g 

Question 13 (b) (iii)

Finding the limit as U   .

 1  k  1  1  2 k  
lim  tan 1  U 2 
  2 kg U   tan  U g  
  lim
U   2 kg g
     
1 
 
2 kg 2


4 kg

– 20 –
Question 13 (c) (i)

We will find PT , then PS , then arrange them as a ratio.

To find PT we group the m people and we will have n  m people remaining. However, we
must also include the group as a single element, so we have  n  m  1! permutations for the
group + everybody else. We then permute the m people in the group, which yields m ! and
therefore PT   n  m  1!m! .

To find PS , we will consider the remaining n  m people. If we put these people next to each
other in a line, there will be n  m  1 ‘gaps’, in which we can place the m people, who are to
be separated. This way, we are guaranteed that they are all separated. Hence to choose
 n  m  1
positions for the m people, we have   . We then permute the m people, which gives
 m 
us m ! and then the remaining n  m people, giving us  n  m ! Therefore
 n  m  1
PS     n  m  !m !
 m 

So putting it as a ratio, we have:

 n  m  1
   n  m  !m !
PS  m 

PT  n  m  1!m !
 n  m  1
 

m 
 n  m  1
 n  m  1!
m ! n  2m  1 !

 n  m  1

 n  m  1!
 n  m  1 m ! n  2m  1!

 n  m !
m ! n  2m  1 !
1
 
 n  m !
m  m  1 ! n  2m  1 !
1 n  m
  
m  m 1 

– 21 –
Question 13 (c) (ii)

n
Suppose that n is even, we can simply substitute m  .
2

PS 1  n  n 2
  
PT n 2  n 2  1 
2  n 2 
  
n  n 2  1
2
 
 n 2 !
n  n 2  1 ! n 2  n 2  1 !
2
 
 n 2 !
n  n 2  1 !
2 n
 
n 2
1

Therefore PS  PT .

Alternatively

You could substitute n  2m :

PS 1  2m  m 
  
PT m  m  1 
1  m 
  
m  m  1
1
 m
m
1

And hence the result.

– 22 –
Question 14 (a) (i)

If we take the xy plane to be along the diagonal of the base, we have the following situation.

However, recall that the base of the vertex lies on the corner of the base, meaning that the
2
vertex is exactly units away from the base. We therefore have the equation of the
2
parabolas being:
2 2
 2 2  2 2
y   x   , 0  x  y   x   ,  x0
 2  2  2  2

To find the value of h in terms of d, we can just use the function. From the diagram, we see
d
that when x  , we have y  h .
2

Using the equation of the parabola on the right, we have:

2
d 2
h    
2 2 
1
 h d 2
2
  
... Since d  2 , we take the negative root 
d  2  2 h
d  2 2 h

 2 1  2h 

– 23 –
Question 14 (a) (ii)

Observe that H is simply the y intercept, so it immediately follows, by letting x  0 , that

1
H .
2

Question 14 (a) (iii)

From the diagonal of the slice, we can deduce the area of the slice.

Since the slice is a square, we can use Pythagoras’ Theorem. Let the side length of the square
be s.

s2  s2  d 2
2s 2  d 2

 
2
2 2h  1

 
2
s2  2h  1

So therefore, the volume of the slice is:

 
2
V  2h  1  h

1
Integrate from 0 to :
2

 
2
V  lim 2h  1  h
 h 0
0  h 1 2
1

 
2 2
 2h  1 dh
0
1
2 1

   2h  2 2h 2  1 dh
0  
1
2 2
 h 2 2 h h h
2

3 0
1
2
4 2
 h2  h h h
3 0

1 4 2 1 1
   
4 3 2 2 2
1

12

– 24 –
Question 14 (b) (i)

We will use the Cosine Rule.

Using the Cosine Rule, we have:

 2 
zk  zk 1  zk  zk  2 zk zk cos 
2 2 2

 n 
 2 
 1  1  2 cos   ...  since zk  1
 n 
 2 
 2  2 cos  
 n 
 2 
zk  zk 1  2  2 cos  
 n 

– 25 –
Alternatively

2 k 2  k  1 
For simplicity, we let x  and y  .
n n

zk  zk 1  cos x  i sin x  cos y  i sin y

 cos x  cos y  i sin x  sin y 

  cos x  cos y    sin x  sin y 

2 2

 cos 2 x  2 cos x cos y  cos 2 y  sin 2 x  2sin x sin y  sin 2 y

 cos 2 x  sin 2 x  cos 2 y  sin 2 y  2  cos x cos y  sin x sin y 

 2k 2  k  1  
 2  2 cos   
 n n 
 2k  2  k  1  
 2  2 cos  
 n 
 2k  2k  2 
 2  2 cos  
 n 
 2 
 2  2 cos  
 n 

Question 14 (b) (ii)

Geometrically, zk  zk 1 is the distance between the endpoints of the two vectors.

So it is the equivalent of the length of one side of the polygon. But we have an n-sided
regular polygon, in which case the perimeter would be Pn  n zk  zk 1 .

– 26 –
Question 14 (b) (iii)

2
Let   , so as n   ,   0 .
n

 2 
lim Pn  lim n 2  2 cos  
n  n 
 n 
2
 lim 2  2 cos 
 0 
2   
 lim 2  2 1  2sin 2   
 0    2 
2  
 lim 2  2  4sin 2  
 0  2
2  
 lim  2sin  

 0
2
4  
 lim  sin  
 0 
2
 
sin  
 lim 4   21
 0   2
 
2
1
 4 
2
 2

Alternatively


Let   , so as n   ,   0 .
n

 2 
lim Pn  lim n 2  2 cos  
n  n 
 n 

 lim 2  2 cos 2

 0


 lim 2 1  cos 2 
 0 


 lim 4sin 2 
 0 

2 sin 
 lim
 0 
 2
– 27 –
Question 14 (b) (iv)

The above result is the perimeter of a circle, since the circle is a limiting case of an n-gon
when n gets infinitely large.

Question 14 (c) (i)

The point D is a common point.

We know that a circle is unique to a set of three distinct points, in this case D, X and Y. So a
circle can indeed be constructed through those points. Since D is a mutual point, CD is also a
mutual tangent.

Alternatively

We could do a relatively simple angle chase:

CDY  DEB (Alternate Segment Theorem)

DEB  DXY (Corresponding Angles, since CW // BE)

Hence CDY  DXY , satisfying the converse of the Alternate Segment Theorem.

Question 14 (c) (ii)

Using the Secant-Chord Theorem with circle DXY, we have:

DC 2  CY  CX

Similarly for circle DBE, we have:

DC 2  CZ  CW

Equate the two ratios:

CY  CX  CZ  CW

Re-arrange the letters to fit the terms of the question:

CY  XC  ZC  CW

Re-arrange the expressions:

CW CY

XC ZC

– 28 –
But note that CW  CX  XW and CY  CZ  ZY . Substitute this in:

CX  XW CZ  ZY

XC ZC
XW ZY
1  1
XC ZC
XW ZY

XC ZC

– 29 –
Question 15 (a) (i)
a

We evaluate I n 1  lim  e x dx .
a 
0
a

I n 1  lim  e  x dx .
a 
0
a
 lim  e  x 
a  0
a
  lim e  x 
a  0

  lim e  e0 
a
a 

   1
1

Question 15 (a) (ii)

a

Shift the subscript up by 1 to get I n 1  lim  x n e x dx .
a 
0

Using Integration by Parts, we have:

u  xn dv  e x dx
du  nx n 1dx v  e  x
a
a 
I n 1  lim  x n e  x  lim n x n 1e  x dx
a  0 a 
0
 lim  a n e  a  nI n
a 

 nI n

Question 15 (a) (iii)

Substitute into itself recursively:

I n 1  nI n
 n  n  1 I n 1
 n  n  1 n  2  I n  2
 ...
 n   n  1   n  2   ...2 1 I1

But I1  1 , and hence I n1  n!

– 30 –
Question 15 (a) (iv)
a

Using the previous integral I n 1  lim  x n e x dx, we make a substitution.
a 
0

dt
Let x   ln t , such that dx   .
t

x  a  t  e a
x  0  t 1

But since in the previous integral, we had a   , we will similarly have t  0 and thus the
new limit of integration will be some value b getting infinitely small.
b
 dt
I n 1  lim    ln t  eln t 
n

b 0
1 t
b
 dt
 lim   1  ln t  t 
n 1 n

b 0
1 t
b

 lim   1  ln t  dt
n 1 n

b 0
1
1

 lim   1  ln t  dt ...  swapping limits uses up a  1
n n

b 0
b
1

  1 lim   ln t  dt
n n

b 0
b
  1 J n
n

So thus we have  1 J n  I n1 . But from (i), we have I n1  nI n , which implies
n

 1 J n  nI n .
n

 1
n
I
Hence n 
Jn n

– 31 –
Alternatively

In
You could work out the actual reduction formula for J n , then find the ratio .
Jn

u   ln t 
n

dv  dt
 ln t 
n 1

du  n  dt vt
t
1

J n  lim   ln t  dt
n

b 0
b
1
n 1 
 lim t  ln t    n  lim   ln t  dt
n 1

b 0  b b 0
 b
1

 lim 0  b  ln b    n  lim   ln t  dt
n n 1

b 0   b 0
 b
1

 n  lim   ln t  dt ...  see Note 
n 1

b 0
b
 nJ n 1
NOTE
Computing J 0 : All people, who attempted this question and
acquired the result, assumed the limit
1
 lim b  ln b   0 without proof.
J 0  lim   ln t  dt
0 n

b 0
b 0
b
1
 However, they were not penalised because the
 lim  dt
b 0
b average would have otherwise been too low.
1 The proof for this, which has been omitted
from this document, uses the Squeeze Law and
So recursively, we acquire J n   1 n!.
n

L’Hopital’s Principle.

Placing this as a ratio with I n :

I n  n  1 !

J n  1n n !

 1  
n
1
 ...  since    
n
= 1
n   1 
n

 

– 32 –
Question 15 (b) (i)

Summing the areas of the upper rectangles, we have

U  f 1 1  f  2  1  f  3  1  ...  f  n  1  1
 f 1  f  2   f  3  ...  f  n  1
1 1 1 1
    ... 
1 2 3 n 1
 H n 1

Observe that:
n

En  H n 1   dx
1
1 x
 H n 1  ln n

1
Adding to both sides, we acquire:
n

1 1
En   H n 1  ln n 
n n
 H n  ln n

Question 15 (b) (ii)

Taking the limit as n   , we have:

1
lim  H n  ln n   lim En 
n  n  n


1
Since for large values of n, the term approaches 0.
n

– 33 –
Question 15 (b) (iii)

We will use Mathematical Induction on n.

Base Case: n  1 .

LHS  S 2 RHS  H1  H 2
 1
r 1
1 2 1
  
2

rr 1 r 1 r r 1 r

1 1 1 1 1
    
1 2 1 2 1
1 1
 
2 2

Thus the base case has been proven.

Inductive Hypothesis: n  k .

S2k  H k  H 2k , for all k  .

Inductive Step: k  k  1 .

Required to prove: S2k 2  H k 1  H 2 k 2

 1
r
2k 2
S2 k  2  
r 1 r
 1  1
r r
2k 2k 2
  
r 1 r r  2 k 1 r
1 1
 S2 k  
2k  1 2k  2
1 1
 H k  H 2k  
2k  1 2  k  1
1 1 1
 H k  H 2k   
2k  1 k  1 2  k  1
 1   1 1 
  Hk     H 2k   
 k 1  2k  1 2k  2 
 H k 1  H 2 k  2
 RHS

Thus, via proof of induction, true for all n  .

– 34 –
Question 15 (b) (iv)

Recall that I r 1  r !

We know that S2n  H n  H 2n .

Using previous parts, we have:

lim  H n  ln n    lim  H 2 n  ln 2n   
n  n 

So using the expression S2n  H n  H 2n , we have:

S2 n   H n  ln n    H 2 n  ln 2n    ln n  ln 2n 
 n 
  H n  ln n    H 2 n  ln 2n   ln  
 2n 
1
  H n  ln n    H 2 n  ln 2n   ln  
2

Taking the limit as n   of both sides:

  1 
lim S2 n  lim  H n  ln n    H 2 n  ln 2n   ln   
n  n 
  2 
1
     ln  
2
1
 ln  
2

Ir
But we observe that the summands are , so:
Jr

 1
r
n
lim S 2 n  lim 
n  n  r
r 1
n
Ir
 lim 
n  Jr
r 1

 lim 
n
 r  1!
n  Jr
r 1
n
r!
 lim 
n  J r 1
r 0

n
r! 1
Hence lim   ln  
n 
r 0 J r 1 2

– 35 –
Question 16 (a)

We first prove a standard result: x  y  x  y .

x  y  x   y
 x  y
 x y

Using the above inequality, we have

 z0  z1  z2  ...  zn    z1  z2  ...  zn   z0  z1  z2  ...  zn  z1  z2  ...  zn

z0  z0  z1  z2  ...  zn  z1  z2  ...  zn
 z0  z1  z2  ...  zn  z1  z2  ...  zn
z0  z1  z2  ...  zn  z0  z1  z2  ...  zn

Question 16 (b) (i) (1)

n
1  x  R  x   1  x   bk x k
k 0
n n
  bk x k  x bk x k
k 0 k 0
n n
 b x  b x
k 0
k
k

k 0
k
k 1

  b0  b1 x  b2 x 2  ...  bn x n    b0 x  b1 x 2  b2 x 3  ...  bn x n 1 

 b0   b1  b0  x   b2  b1  x 2  ...   bn  bn 1  x n  bn x n 1
 b0   b1  b0  x   b2  b1  x 2  ...   bn  bn 1  x n  bn x n 1
n 1
 b0  b1  b0 x  b2  b1 x  ...  bn  bn 1 x  bn x
2 n

 b0   b0  b1  x   b1  b2  x  ...   bn 1  bn  x  bn x ...  since b n  bn 1 

2 n n 1

n
 b0    bk 1  bk  x  bn x
k n 1

k 1

– 36 –
Question 16 (b) (i) (2)

When x  1 , then the inequality becomes

n
1  x  R  x   b0    bk 1  bk  x
n 1
 bn x
k

k 1
n
 b0    bk 1  bk   bn
k 1

0 ...  since we have a Telescoping Sum 

Question 16 (b) (i) (3)

Since 1  x  R  x   0 when x  1 , we have R  x   0 . This means that since R  x   0

when x  1 , any root  of R  x  must satisfy   1 .

Question 16 (b) (ii) (1)

Since  is a root of S  x  , we have S     c0  c1  c2  2  ...  cn  n  0 .

1
We now evaluate T   .
 

1 n 1
T     cn  k k
   k 0 
n
  cn  k   k
k 0
n
1
   cn  k  n  k
 n
k 0

  cn  n  cn 1 n 1  ...  c1  c0 

1

 n

0

– 37 –
Question 16 (b) (ii) (2)

Likewise for R  x  , we acquire:

n
1  x  T  x   cn   ck  ck 1
nk n 1
x  c0 x
k 1

Note that the exponent is n  k and that we have ck  ck 1 as opposed to ck 1  ck like before.
This is due to the nature of the coefficients (in a sense ‘reversed’) of T  x  .

Again, similarly to (i), having x  1 yields us another Telescoping Sum, which results in
1  x  T  x   0 . Suppose  is some root of T  x  . We then have   1 , similarly to (i).

1 1
However, recall that   due to (b) (ii) (1), hence  1 and therefore   1 .
 

Question 16 (c) (i)

ak 1
We know that A  , for all 1  k  n , since A is defined to be the minimum.
ak

Re-arranging, we have:

Ak a
k 1
 k 1
A ak
ak Ak  ak 1 Ak 1

And hence we have (noting that all the values are positive since ak  0  A  0 ) :

0  an An  an1 An1  ...  a2 A2  a1 A  a0

– 38 –
Question 16 (c) (ii)

If we substitute the expression bk  ak Ak into the inequality, we have

0  bn  bn1  ...  b1  b0

Since P    0 , then

a 
k 0
k
k
0
n
k
 bk
k 0 Ak
0

 
n k


k 0
bk    0
 A
 
R   0
 A

The claim, that the polynomial R appears here, is valid because of the inequality with the
coefficients.


Hence is a root of R  x  .
A

Question 16 (c) (iii)

Similarly to (c) (ii), we have ak 1Bk 1  ak Bk and hence 0  a0  a1B  ...  an1Bn1  an Bn .

Using the substitution ck  ak B k , we have 0  c0  c1  ...  cn1  cn . Hence

a 
k 0
k
k
0
n
k
 ck
k 0 Bk
0

 
n k


k 0
ck    0
B
 
S 0
B

The claim, that we have the polynomial S, is valid here because of the inequality with the
coefficients.


And hence is a root of S  x  .
B
– 39 –
Question 16 (d)

 
Since   and   1 , we have  1 and hence   B .
B B

 
Similarly, since   and   1 , we have  1 and hence   A .
A A

Therefore, we obtain the inequality A    B .

– 40 –