1.

Let Pn be the statement

n ( n + 1)
(1 + x )(1 + 2 x )(1 + 3x ) (1 + nx )  1 + x
2
for all positive integers n.
When n = 1,
LHS = 1 + x
(1)(2)
RHS = 1 + x = 1 + x  (or =) RHS
2
Thus P1 is true.
Suppose Pk is true for some positive integer k. Then
LHS of Pk +1
= (1 + x )(1 + 2 x )(1 + 3 x ) (1 + kx ) (1 + ( k + 1) x )
 k ( k + 1) 
 1 + x  (1 + ( k + 1) x )
 2 
k ( k + 1)  k ( k + 1) 
= 1 + ( k + 1) x + x+ x  ( ( k + 1) x )
2  2 
 k 1
= 1 + ( k + 1) 1 +  x + k ( k + 1) x 2
2

 2 2
( k + 1)( k + 2 ) x  1 k k + 1 2 x 2  0 
 1+  ( ) 
2  2 
= RHS of Pk +1
Thus Pk is true implies that Pk +1 is true.
Since P1 is true and Pk is true implies that Pk +1 is true, Pn is true for all positive
integers n by mathematical induction.
n
 k

k =1
ln 1 + 
 n
 1  2  3  n
= ln 1 +  + ln 1 +  + ln 1 +  + + ln 1 + 
 n  n  n  n
 1  2  3   n 
= ln  1 +  1 +   1 +   1 +  
 n  n  n   n 
 n ( n + 1) 1  1
 ln 1 +   (use x = in first part)
 2 n n
 n +1 
= ln 1 + 
 2 
 n+3
= ln  
 2 
= ln ( n + 3) − ln 2 (shown)
2. 0 0
(a) det( A)  0  A is invertible but   is non- invertible.
0 0
 OR
0 0 0 0
det    0   0 0  is not in the set.
0 0  
Since the set does not contain the zero matrix, it is not a subspace of
M 22 ( ) .

(b) Let U denote the set of all 2  2 matrices B such that BT = −B .

T
0 0  0 0  0 0
  =  = − 
0 0  0 0  0 0
0 0
Hence    U and U is non-empty.
0 0

Consider any two elements B , C  U .

( B + C) = BT + CT = −B + ( −C) = − ( B + C )
T

 B + C  U (Closure under addition)

Consider any element B  U and k  .
( kB ) = kBT = k ( −B ) = − ( kB )
T

 kB  U (Closure under scalar multiplication)

Hence U is a subspace of M 22 ( ).

3. dx
x = sec   = sec  tan  .
d
dy dy dx dy
= = sec tan  .
d dx d dx
d 2 y d 2 y  dx 
 sec  tan  + ( sec  tan  + sec  )
dy
= 2 2 3

d 2
dx  d   dx
d2 y
= 2 sec 2  tan 2  + ( sec  tan 2  + sec3  )
dy
dx dx
2
= 2 sec 2  ( sec 2  − 1) + sec  ( sec 2  − 1 + sec 2  )
d y dy
dx dx
2
= 2 ( sec 4  − sec 2  ) + sec  ( 2sec 2  − 1)
d y dy
dx dx
2
= ( x 4 − x 2 ) 2 + x ( 2 x 2 − 1)
d y dy
dx dx
1d y 
2 2

 2 = x −x
x  d 
( 3
) d y
dx
(
2 )
+ 2x2 −1
dy
dx
.

Substituting into the differential equation in x and y:

 1  d 2 y  ky 2
 + =
x  d 2  x x 3
d2 y 2
+ ky = 2
d 2
x
2
d y 2
+ ky =
d 2
sec 2 
d2 y
+ ky = 2 cos 2  (shown)
d 2

Characteristic equation:
m2 + k = 0
m =  ki ( k  0)

Complementary function:
yc = A cos k + B sin k
= A cos k ( sec−1 x ) + B sin k ( sec −1 x ) .

Since 2cos  = cos 2 + 1 , particular integral:

2

y p = C cos 2 + D sin 2 + E
yp = −2C sin 2 + 2 D cos 2
yp = −4C cos 2 − 4 D sin 2 .

Substituting into the new differential equation:

−4C cos 2 − 4 D sin 2 + k ( C cos 2 + D sin 2 + E ) = cos 2 + 1
C ( k − 4 ) cos 2 + D ( k − 4 ) sin 2 + kE = cos 2 + 1
Comparing coefficients:
kE = 1 C ( k − 4) = 1 and D ( k − 4) = 0
E=
1
, C=
1
,
D=0 ( k  4).
k k −4
1 1
 yp = cos 2 + .
k −4 k
1 1
Thus, y = A cos k + B sin k + cos 2 + .
k −4 k

Now, cos 2 = 2 cos 2  − 1

2
= −1
sec 2 
2 2 − x2
= 2 −1 = .
x x2

Thus, the general solution is

 2 − x2
y = A cos k ( sec −1 x ) + B sin k ( sec −1 x ) +
1
+
( k − 4) x k
2

where A and B are arbitrary constants.

4. (i) Let N be the foot of perpendicular from M to the directrix x = − p .

M
N
θ

F (sun)
MF = eMN
r = e (Directrix:
p − (−r cos  ) )

x = − p
r − re cos = ep
ep
r = (shown, where k = ep)
1 − e cos 

(ii) Let the distance from the sun to the aphelion and that from the sun to the
perihelion be d1 and d 2 respectively.
k k
Then d1 =  d1 = − − − (1)
1 − e cos 0 1− e
k k
and d 2 =  d2 = − − − (2)
1 − e cos  1+ e
d 61 1 + e
(1)  (2) : 1 = =
d 2 50 1 − e
 61 − 61e = 50 + 50e
11
e= (shown)
111

11
If the e becomes smaller than but larger than 0, the orbit will get more
111
circular.

61 k
(iii) d1 = (3.04) =
111 1− e
 k = 1.51121316 = 1.51 (to 3 sf)

k dr −k ( e sin  )
r=  =
1 − e cos  d (1 − e cos  )2
Distance travelled in one complete orbit
 2 2
 dr 
=
 r +  d
2

0  d 
2
  k ( e sin  ) 
2


2
 k
=   +  d
 1 − e cos    (1 − e cos  ) 
2
0

= 9.565819311 AU

 the orbital period of planet M

9.565819311
=
5.06
= 1.89 years
5. (i)
z 7 + 128 = 0
z 7 = −128
z 7 = 128ei( 2 k +1) π , k 
i ( 2 k7+1 ) π
z = 2e , k = 0, 1, 2, 3, 4, 5, 6
i 7π i 3π i 5π i 9π i 11π i 13π
= 2e , 2e , 2e 7 , 2eiπ , 2e 7 , 2e
7 7
, 2e 7

(ii), (iv)

(iii)
Considering sum of roots of the equation z 7 + 128 = 0 and taking real part,
 π 3π 5π 9π 11π 13π
cos + cos + cos + cos π + cos + cos + cos =0
7 7 7 7 7 7
π 3π 5π 5π 3π π
cos + cos + cos + (−1) + cos + cos + cos = 0
7 7 7 7 7 7
 π 3π 5π 
2  cos + cos + cos  − 1 = 0
 7 7 7 
π 3π 5π 1
cos + cos + cos =
7 7 7 2
 π  3π   5π  1
1 − 2sin 2   + 1 − 2sin 2   + 1 − 2sin 2   =
 14   14   14  2
  π  3π   5π   1
3 − 2 sin 2   + sin 2   + sin 2    =
  14   14   14   2
 π  3π   5π  3 − 12 5
 sin 2   + sin 2   + sin 2   = =
 14   14   14  2 4
(iv)
Area
1 2   2π  π 1 π
= ( 2 )  sin  3   + sin  or  4  2sin
2   7  7 2 7
π
= 4sin or 1.74 units 2
7

(v)

 ( z − z )( z − z ) 
arg  1 2 3 4 
 ( z3 − z2 )( z1 − z4 ) 
= ( arg ( z1 − z2 ) − arg ( z3 − z2 ) ) + ( arg ( z3 − z4 ) − arg ( z1 − z4 ) )
 This argument represents the sum of two opposite angles in a cyclic quadrilateral
(or equivalently angles in opposite segments) and thus they must add up to 
radians (shown).

(Note that since z1 , z2 , z3 and z4 are any 4 distinct roots, the quadrilateral shown
above is not unique.)
6. (i) H 0 : The colour preference is independent of gender
H1 : The colour preference is dependent on gender
Under H 0 , the expected frequencies are
White Green Black Blue Red Others
Male 92 75.3 65.3 53.3 84 30
Female 46 37.7 32.7 26.7 42 15
Degree of freedom = ( 6 − 1)  ( 2 − 1) = 5
From GC,  2 test = 14.02682385 and p-value = 0.0154399223
Since p-value < 0.05, we reject H 0 at the 5% significance level and conclude that
there is sufficient evidence to claim that preferences for colours of gaming
keyboards differ between male and female gamers.

(ii) The table of contributions are as follows:

White Green Black Blue Red Others
M 0.043 0.00147 2.08 0.0020 2.01 0.533
5 8
F 0.087 0.00295 4.17 0.0041 4.02 1.07
0 7
The largest contributions are from the “Black” and “Red” columns, where there is
a significantly larger proportion of males who prefer black colour and a larger
proportion of females who prefer red colour.
Specifically, more male prefer black while fewer male prefer red than expected,
and more female prefer red while fewer female prefer black than expected.
7. (i) To carry out a t-test on the data will require the condition that the scores (or
difference between the scores) to be normally distributed. This condition might not
be met by the data and hence it is not appropriate to use the t-test.

(ii) Let X = score after module – score before module

Assume that the distribution of the population of the differences is symmetrical.
H 0 : Population median of X is 0
H1 : Population median of X is greater than 0
A B C D E F G H I J K L
1st 71 70 65 42 74 81 56 69 57 77 81 55
2 n 51 72 46 93 95 31 81 84 88 67 70 82
d

X −20 2 −19 51 21 −50 25 15 31 −10 −11 27

R 6 1 5 12 7 11 8 4 10 2 3 9
From table, the sum of negative ranks, Q = 27 .
From MF26, at 1% level of significance with n = 12 ,
One-tail critical value of T  9 ( or  69 )
 Since 27 falls beyond the critical region, we do not reject H 0 and conclude at 1%
level of significance that there is insufficient evidence to claim that the population
median for the scores in the writing test has improved (i.e. insufficient evidence to
conclude that the module is effective).
1
8.
For a p.d.f., 
k
(i)  dx = 1
0 x − x + 1
2

1

1
 k dx = 1

( )
2
2
 3 
0 x − 2 +  2 
1
 
1
  x − 1 
 2 −1  2  = 1
 k tan  

3  3 
 2  0
1
2k  −1  2 x − 1  
  tan   = 1
3  3 0
2k  −1  1  −1  −1  
  tan   − tan   = 1
3  3  3 
2k  π π 
 + =1
3  6 6 
3 3
 k=
2π

(ii) From symmetry, Mean = Median = Mode = 0.5

x

3 3 1
(iii) For 0  x  1 , P( X  x ) = dt
2π 
( )
2
2
 3 
0 t − 1 2 +  2 
 
x
3  2t − 1  
=  tan −1  
π  3 0
3  2x −1  1
= tan −1  +
π  3  2
0 for x  0,

3  2x −1  1
Hence F( x ) =  tan −1  + for 0  x  1,
π  3  2
1 for x  1.

3 3  50 tan −1 x
1

(iv) E(Y ) = E( 50 tan −1 X ) =  dx

2 π 0 x 2 − x + 1
= 22.03293348  22.0 (3 s.f.)
   1 
(v) P(Y  25 ) = P( 50 tan −1 X  25 ) = P X  tan   
  2 
3 3 1 1
=  dx = 0.4489963784  0.449 (3 s.f.)
2π tan ( 12 ) x − x + 1
2

9. (i) For the first design, a 2-sample t-test is more appropriate and the benefit is
that it is faster to implement since the cameras only need to go through a single
round of data collection.
For the second design, a paired-sample t-test is more appropriate and the
benefit is that we can be more confident that any statistical difference is due to the
change in batteries since all other factors would be kept relatively constant.

(ii) We need to assume that the battery life for both battery models follows two
independent normal distributions with a common variance.

From data,
( x − x ) + ( y − y )
2 2
961 + 1369
sp 2
= = = 179.23
7+8−2 13

To test H0 : Y =  X vs H1 : Y   X
Y −X
Under H 0 , T = ~ t ( n X + nY − 2 ) = t (13)
1 1
Sp +
nX nY

From GC, t = 2.1649 and p-value = 0.0248 .

For such a p-value, we can conclude that there is sufficient evidence at the 10%
and 5% levels of significance but not at the 1% level of significance that the new
battery model has a longer battery life, making the evidence relatively strong.

(iii) From GC, the 99% CI is (783.7, 818.3).

10. (i) We need to assume that

(1) the arrivals of customers occur independently, and
(2) the average rate of customer arrivals over any time period remains constant.

(ii) Assume that the number of customers arriving and the number of food
delivery orders received are independent.
Let S be the random variable denoting the sum of the number of customers arriving
and the number of food delivery orders received in a 10-minute interval.
Then S ~ Po ( 2 + 5) = Po ( 7 ) .
P( S  10) = 0.901479206  0.901 (3 s.f.)

(iii) The conditions necessary are

(1) the event that each successive interval is busy is independent of all other
intervals,and
(2) the probability of each interval being busy is constant.

The conditions are addressed by the independent nature of the Poisson distribution.

(iv) P( S  10 ) = 0.098521
From (b), B ~ Geo ( 0.098521)
1
E( B ) = = 10.1501415  10.2 (3 s.f.)
0.098521
1 − 0.098521
Var ( B ) = = 92.87523101  92.9 (3 s.f.)
0.0985212

(v) P( B  m )  0.25  P( B  m )  0.75

From GC, least m = 14