Suggested Solution for Pure Mathematics 2008 (I) Y.K.

Ng (last update: 10/4/08)

Paper I 7x + 9 A B C
2. (a) Let = + + .
m n x( x + 1)( x + 3) x x + 1 x + 3
1. (a) (1 + x) m = ∑ m Ck x k and (1 + x) n = ∑ n Cs x s .
k =0 s =0 Then 7 x + 9 = A( x + 1)( x + 3) + Bx( x + 3) + Cx ( x + 1) for all x.
m n Put x = 0 , −1 and −3 , we have A = 3 , B = −1 and C = −2
∴ (1 + x) m + n = (1 + x) m (1 + x) n = ( ∑ m C k x k )( ∑ n C s x s )
k =0 s =0 7x + 9 3 1 2
i.e. = − −
m n x( x + 1)( x + 3) x x + 1 x + 3
= ∑ ∑ m Ck ⋅n Cs x k +s n n
k =0 s =0 7k + 9 3 1 2
r
(b) ∑ k (k + 1)(k + 3) = ∑ ( k − k + 1 − k + 3 )
Compare coefficient of x : k =1 k =1

m n
3 3 2 2
m+ n C r = ∑ m C k ⋅ n C r −k = ∑ [( −
k +1
)+( −
k +1 k + 3
)]
k =0 k =1 k

0 ≤ k ≤ m 0 ≤ k ≤ m 0 ≤ k ≤ m n
3 3 2 2 2 2
[Note that:  gives  ,  = ∑ [( − )+( − )+( − )]
0 ≤ s ≤ n 0 ≤ r − k ≤ n r − n ≤ k ≤ r k =1 k k +1 k +1 k + 2 k +2 k +3

i.e. 0 ≤ k ≤ m as r = m , m + 1 , …, n.] 1 1 1 1 1
= 3(1 − ) + 2( − ) + 2( + )
99 99 n +1 2 n+2 3 n+3
∑ 99 C k ⋅100 C k +1 ∑ 99 C k ⋅100 C 99− k 14 3 2 2
k =0 k =0
(b) 99
= 99
= − − −
3 n +1 n + 2 n + 3
∑ 99 C k ⋅101 C k ∑ 99 C k ⋅101 C101− k
k =0 k =0 4
7k + 9 362
(c) Note that ∑ k (k + 1)(k + 3) = 105
99 +100 C 99 k =1
=
99 +101 C101 ∞
7k + 9 n
7k + 9
199 !
∴ ∑ k (k + 1)(k + 3) = nlim
→∞
∑ k (k + 1)(k + 3)
k =5 k =5

= 99 !100 ! n 4
7k + 9 7k + 9
200 ! = lim ( ∑ −∑ )
101! 99 !
n →∞ k =1 k ( k + 1)(k + 3) k =1 k ( k + 1)( k + 3)

101 14 3 2 2 362
= = lim ( − − − − )
200 n→∞ 3 n + 1 n + 2 n + 3 105

128
=
105

Page 1
Suggested Solution for Pure Mathematics 2008 (I) Y.K.Ng (last update: 10/4/08)
a b c  −1
3 3 3  − 3 
3. (a) a + b + c − 3abc = c a b  cos 120 o o 
− sin 120   2 2 
4. (a) A =  =
c b a
 sin 120
o
cos 120 o   3 −1 
 
 2 2 
a+b+c b c
= a+b+c a b ( C1 + C 2 + C 3 ) (b) (i) Let P = (x, y).
a+b+c b a
x  4 
Q A  =  
 
=0 (Q a+b+c = 0 )  y 2 3 
(b) Put a = x − 2 2 + 5 3 , b = −3x − 6 2 + 7 3 and c = 2 x + 8 2 − 12 3
 −1 3 
4  
2 
such that a + b + c = 0 x 4   1 
∴   = A −1  =  2 = 
2 3  − 1  2 3   − 3 3 

Then, 0= ( x − 2 2 + 5 3 ) 3 − (3x + 6 2 − 7 3 ) 3 + 8( x + 4 2 − 6 3 ) 3  y   3
− 
 2 2 
= a3 + b3 + c3
= 3abc i.e. P = (1, − 3 3 )
= 3( x − 2 2 + 5 3 )(−3x − 6 2 + 7 3 )(2 x + 8 2 − 12 3 ) (ii) Rotation in the Cartesian plane anticlockwise about the origin by

7 3 120n o .
i.e. x = 2 2 −5 3 , −2 2 + or − 4 2 + 6 3
3 (iii) Note that the distance between origin and P equals the distance
between origin and the image of P under the rotation by T n .

Now, OP = 12 + (3 3 ) 2 = 28

OQ = 3 2 + (5 3 ) 2 = 84 ≠ 28 = OP, where Q = (3, 5 3 )

i.e. There is no transformation such that P is transformed to Q.

Page 2
Suggested Solution for Pure Mathematics 2008 (I) Y.K.Ng (last update: 10/4/08)
5. (a) Put z = x + iy . 6. (a) (i) Consider the two set of numbers: {1, 1, 1, 1} and { α, β, γ, δ }.
3 | z − 2i |=| z + 8 − 2i | Applying Cauchy-Schwarz’s inequality, we have
3 2 | x + ( y − 2)i | 2 =| ( x + 8) + ( y − 2)i | 2 (1 ⋅ α + 1 ⋅ β + 1 ⋅ γ + 1 ⋅ δ) 2 ≤ (12 + 12 + 12 + 12 )(α 2 + β 2 + γ 2 + δ 2 )
9[ x 2 + ( y − 2) 2 ] = ( x + 8) 2 + ( y − 2) 2 i.e. (α + β + γ + δ) 2 ≤ 4(α 2 + β 2 + γ 2 + δ 2 )
i.e. x 2 + y 2 − 2x − 4 y − 4 = 0 α β γ δ
the equality holds iff = = = iff α = β = γ = δ
which is a circle with centre at 1+ 2i and 1 1 1 1
1 1 1 1
radius 1 2 + 2 2 − ( −4 ) = 3 (ii) Consider the two set of numbers: { α 2 , β 2 , γ 2 , δ 2 } and

(b) The longest distance 3 3 3 3

= distance between P and centre of circle + radius of the circle { α 2 , β 2 , γ 2 , δ 2 }.

= | (−7 + 17i ) − (1 + 2i ) | +3 Applying Cauchy-Schwarz’s inequality, we have

1 3 1 3 1 3 1 3
= 8 2 + 15 2 + 3 (α 2 ⋅α 2 +β2 ⋅β 2 +γ2 ⋅γ 2 +δ2 ⋅δ 2 )2

= 20 1 1 1 1
≤ [(α 2 ) 2 + (β 2 ) 2 + (γ 2 ) 2 + δ 2 ) 2 ]×

3 3 3 3
[(α 2 ) 2 + (β 2 ) 2 + (γ 2 )2 + δ 2 )2 ]

i.e. (α 2 + β 2 + γ 2 + δ 2 ) 2 ≤ (α + β + γ + δ)(α 3 + β 3 + γ 3 + δ 3 )
α3/ 2 β3/ 2 γ3/ 2 δ3 / 2
The equality holds iff = = =
α1 / 2 β1 / 2 γ1/ 2 δ1 / 2
iff α = β = γ = δ

Page 3
Suggested Solution for Pure Mathematics 2008 (I) Y.K.Ng (last update: 10/4/08)
7. (a) (i) Coefficient determinant D
2 2 2 2 2
(b) Now (α + β + γ + δ) ≤ 4(α + β + γ + δ ) 1 a + 2 a +1
3 3 3 3
(α + β + γ + δ)(α + β + γ + δ ) ≥ (α + β + γ + δ )2 2 2 2 2 = 1 −3 − 1 = −a 2 + 4
Divides two inequalities, 3 − 2 a −1

α+β+ γ +δ 4( α 2 + β 2 + γ 2 + δ 2 ) (E) has unique solution ⇔ D ≠ 0 ⇔ a 2 ≠ 4

≤
α 3 + β3 + γ 3 + δ3 (α 2 + β 2 + γ 2 + δ 2 ) 2 In this case,
α + β + γ + δ α 3 + β3 + γ 3 + δ 3 1 a + 2 a +1
i.e. ≤ 2
4 α + β2 + γ 2 + δ2 b −3 −1
The equality holds Dx 1 −2 a −1 2 − a − a 2 b − 3ab
x= = 2
=
iff the two equalities of (a)(i) and (a)(ii) hold D −a +4 − a2 + 4
iff α = β = γ = δ
1 1 a +1
1 b −1
Dy 3 1 a −1 − 2b(a + 2) 2b
y= = 2
= 2
=
D −a +4 −a +4 a−2

1 a+2 1
1 −3 b
Dz 3 −2 1 2 − a + 8b + 3ab
z= = 2
=
D −a +4 − a2 + 4

Page 4
Suggested Solution for Pure Mathematics 2008 (I) Y.K.Ng (last update: 10/4/08)
(ii) (1) When a = 2 , D = 0 (b) The given system is the same as (E) by putting a = 2 and b = 0 .

 x + 4 y + 3z = 1 − − − < 1 > 3 − 5t 1 − 4t
 By (a)(ii)(1), the solution is x = , y= , z = t for t ∈ ℜ .
(E):  x − 3 y − z = b − −− < 2 > 7 7
3 x − 2 y + z = 1 − −− < 3 >
 Then, 2 x 2 + 15 y 2 − 10 z 2
On solving <1> and <3>, we have 3 − 5t 2 1 − 4t 2
= 2( ) + 15( ) − 10t 2
3 − 5z 1 − 4z 7 7
x= and y =
7 7 1
= (−200t 2 − 180t + 33)
For (E) to be consistent, the above solution satisfy <2>. 49

3 − 5z 1 − 4z 1 9 147
∴ b = x − 3y − z = − 3⋅ −z=0 = [−200(t − ) 2 + ]
7 7 49 20 2
Solution set in this case 1 147 3
≤ ⋅ =
3 − 5t 1 − 4t 49 2 2
= {( , , t ): t ∈ ℜ }
7 7 3
i.e. maximum value of the expression = .
(2) When a = −2 , D = 0 2

x − z =1 −−− < 4 >


(E):  x − 3 y − z = b − −− < 5 >
3 x − 2 y − 3z = 1 − −− < 6 >

On solving <4> and <6>, we have
x = 1 + z and y = 1
For (E) to be consistent, the above solution satisfy <5>.
b = x − 3 y − z = (1 + z ) − 3 ⋅1 − z = −2
Solution set in this case
= {( 1 + t , 1, t ): t ∈ ℜ }

Page 5
Suggested Solution for Pure Mathematics 2008 (I) Y.K.Ng (last update: 10/4/08)
 p 2 + qr 0  2 3 
8. (a) (i) M 2 =   (b) (i) A =   + 4 I
 0 p + qr 
2
 −1 − 2
(ii) When p 2 + qr = 1 , M 2 = I . (ii) Put p = 2 , q = 3 r = −1 such that p 2 + qr = 1 .
For n = 1, ( M + sI ) 2 = M 2 + 2sM + s 2 I Then A = M + 4 I where M 2 = I .
= 2sM + ( s 2 + 1) I A 800 = ( M + 4 I ) 2( 400)

( s + 1) 2 − ( s − 1) 2 ( s + 1) 2 + ( s − 1) 2 5 800 − 3800 5 800 + 3800

= M+ I = M+ I (by (a)(ii))
2 2 2 2
∴ the statement is true for n = 1. 5 800 − 3800 2 3  5 800 + 3800 1 0
=   +  
( s + 1) 2 k − ( s − 1) 2 k ( s + 1) 2 k + ( s − 1) 2 k
2  −1 − 2 2 0 1
Assume ( M + sI ) 2 k = M+ I
2 2
 3 ⋅ 5 800 − 3800 3 ⋅ 5 800 − 3 ⋅ 3800 
for some positive integer k.  
= 2 2 
Consider  − 5 800 + 3800 − 5 800 + 3 ⋅ 3800 
 
 2 2 
( M + sI ) 2( k +1) = ( M + sI ) 2 k ( M + sI ) 2
 ( s + 1) 2 − ( s − 1) 2 ( s + 1) 2 + ( s − 1) 2  (iii) Note that | A |= 15 , so | A 800 |=| A |800 = 15 800 .
=

M+ (
I  × 2sM + ( s 2 + 1) I

)
 2 2  ( A −1 ) 800 = ( A 800 ) −1
 ( s + 1) 2 k − ( s − 1) 2 k ( s + 1) 2 k + ( s − 1) 2 k 
= M  ( s 2 + 1) + 2s   − 5 800 + 3 ⋅ 3800 − 3 ⋅ 5 800 + 3 ⋅ 3800 
 2 2   
  1  2 2 
=
  15 800  + 5 800 − 3800 3 ⋅ 5 800 − 3800 
( s + 1) 2 k + ( s − 1) 2 k ( s + 1) 2 k − ( s − 1) 2 k  
+ I  ( s 2 + 1) + 2s 
 
 2 2  2 2
 
 ( s + 1) 2 k ( s 2 + 1 + 2s ) ( s − 1) 2 k ( s 2 + 1 − 2s ) 
= M − 
 2 2 
 
 ( s + 1) 2 k ( s 2 + 1 + 2s ) ( s − 1) 2 k ( s 2 + 1 − 2s ) 
+ I + 
 2 2 
 
( s + 1) 2( k +1) − ( s − 1) 2( k +1) ( s + 1) 2( k +1) + ( s − 1) 2( k +1)
= M+ I
2 2
∴ By induction, the statement is true for all n ∈ Z+.

Page 6
Suggested Solution for Pure Mathematics 2008 (I) Y.K.Ng (last update: 10/4/08)
xn 2 y n + xn y n 2 xn 2 + yn 2 (c) By (b), we have
9. (a) x n +1 − y n +1 = 2 2
−
xn + yn xn + yn 2 = x1 ≤ x 2 ≤ ... ≤ x n ≤ y n ≤ y n −1 ≤ ... ≤ y 2 ≤ y1 = 8

( x n 2 y n + x n y n 2 )( x n + y n ) − ( x n 2 + y n 2 ) 2 { x n } is increasing and bounded above by 8, so lim x n = L1 exists.

= n →∞
( x n 2 + y n 2 )( x n + y n )
xn 3 y n + 2xn 2 y n 2 + xn yn 3 − (xn 4 + 2xn 2 yn 2 + yn 4 ) { y n } is decreasing and bounded below by 8, so lim y n = L 2 exists.
= n →∞
( x n 2 + y n 2 )( x n + y n )
(xn 3 yn − xn 4 ) + (xn y n 3 − yn 4 ) xn 2 + yn 2
= Taking limit on both sides of y n +1 = , we get
( x n 2 + y n 2 )( x n + y n ) xn + yn

− xn 3 (xn − yn ) + yn 3 (xn − y n ) L 2 ( L1 + L 2 ) = L1 2 + L 2 2
=
( x n 2 + y n 2 )( x n + y n ) L1 ( L1 − L 2 ) = 0
− ( x n 3 − y n 3 )( x n − y n ) L1 = L2 only (since x n ≥ 2 , L1 = lim x n ≥ 2 )
= n →∞
( x n 2 + y n 2 )( x n + y n )
(b) (i) First of all, it is trivial that x n , y n >0 for all positive integer n. x n −1 2 y n −1 + x n −1 y n −1 2 x n −1 2 + y n −1 2
(d) xn yn = ⋅ = x n −1 y n −1 = ... = x1 y1 = 16
x n −1 2 + y n −1 2 x n −1 + y n −1
For n = 1, x1 = 2 ≤ 8 = y1 Taking limits on both sides, L1 L 2 = 16
Assume x k ≤ y k for some positive integer k. i.e. L1 = L 2 = 4 (note: L1 , L 2 > 0 )

− ( x k 3 − y k 3 )( x k − y k )
x k +1 − y k +1 = ≤0
( x k 2 + y k 2 )( x k + y k )
By induction, x n ≤ y n for all positive integer n.

xn 2 yn + xn yn 2 xn 2 ( yn − xn )
(ii) x n +1 − x n = − xn = ≥0
xn 2 + yn 2 xn 2 + yn 2

xn 2 + y n 2 x ( y − xn )
(iii) y n − y n +1 = y n − = n n ≥0
xn + y n xn + yn

Page 7
Suggested Solution for Pure Mathematics 2008 (I) Y.K.Ng (last update: 10/4/08)
k (b) Substitute x = 1 into p( x) = p( x − 1) + x 100
10. (a) Let p( x) = ∑ a r x r , where ak ≠ 0
r =0 1 = p (1) = p(0) + 1 , i.e. p ( 0) = 0
Q p( x) = p( x − 1) + x 100 Substitute x = 0 into p( x) = p( x − 1) + x 100
k k 0 = p(0) = p(−1) + 0 , i.e. p(−1) = 0
∴ ∑ a r x r = ∑ a r ( x − 1) r + x 100
r =0 r =0 (c) p( x) = p( x − 1) + x 100
Case (i) 0 ≤ k < 100 –) p(− x) = p(− x − 1) + (− x)100
Compare coefficient of x 100 , we have 0 = 1, impossible. p( x) − p(− x ) = p( x − 1) − p (− x − 1)
Case (ii) k = 100 i.e. p( x) + p (− x − 1) = p ( x − 1) + p(− x)
Compare coefficient of x 100 , we have a100 = a100 + 1 , which For n = 1, p(0) + p(0 − 1) = 0 + 0 = 0
is impossible. Assume p(k ) + p (−k − 1) = 0 for some positive integer k.
Case (iii) k ≥ 102 Consider p(k + 1) + p (−(k + 1) − 1)
k −1 = p ((k + 1) − 1) + p(−(k + 1)) (put x = k + 1 in (c))
Compare coefficient of x , we have
a k −1 = ka k + a k −1 , = p (k ) + p(−k − 1)
ka k = 0 which is impossible since k , a k ≠ 0 =0 (by assumption)
Therefore, k = 101 By induction, the statement is true for all positive integers n.
101 101 (d) Let F ( x) = p ( x) + p(− x − 1) .
By comparing coefficient of x 100 in ∑ a r x r = ∑ a r ( x − 1) r + x100 ,
r =0 r =0 * deg F ( x) = k = 101 by (a)

1 * F (1) = F (2) = ... = F (102) = 0 by (c)

we have a100 = 101a101 + a100 + 1 , a101 =
101 Therefore, F (x) has more than 101 distinct roots but with degree 101.

1 i.e. F ( x) ≡ 0
i.e. the coefficient of x k is .
101 i.e. p( x) + p(− x − 1) = 0 for all x ∈ ℜ

1 1 1 1
(e) Put x = − in (d), 0 = p(− ) + p(− ) , ∴ p(− ) = 0
2 2 2 2
Further, p(0) = p(−1) = 0
Therefore x , x + 1 2 x + 1 are the factors of p(x) .
i.e. p(x) is divisible by x( x + 1)(2 x + 1)

Page 8
Suggested Solution for Pure Mathematics 2008 (I) Y.K.Ng (last update: 10/4/08)
11. (a) (i) Suppose the converse that there exists integers p and q (≠ 0) where p (b) Existence
p For n = 1, take a1 = 3 , b1 = 1
and q are relatively prime, such that 10 = .
q then, (3 + 10 )1 = 3 + 10 = a1 + b1 10
Then p 2 = 10q 2 , and (3 − 10 )1 = 3 − 10 = a1 − b1 10 .
∴ p is divisible by 10, (since p and q are relatively prime) ∴ The statement is true for n = 1.
say, p = 10k for some integer k. Assume there exists positive integers a k and bk such that
∴ 10q 2 = p 2 = 100k 2 , q 2 = 10k 2 , (3 + 10 ) k = a k + bk 10 and (3 − 10 ) k = a k − bk 10
∴ q is divisible by 10. (Q q contains no factor of p) for some positive integer k.
Take a k +1 = 3a k + 10bk ∈ Z+ and bk +1 = a k + 3bk ∈ Z+.
Now, both p and q have the common factor 10, which (3 + 10 ) k +1 = (3 + 10 ) k (3 + 10 )
contradicts to the supposition that they are relatively prime. = (a k + bk 10 )(3 + 10 )
i.e. 10 is an irrational number. = (3a k + 10bk ) + (a k + 3bk ) 10
= a k +1 + bk +1 10
k +1
(ii) Suppose the converse that there exists non-zero integers m and n such (3 − 10 ) = (3 − 10 ) k (3 − 10 )
that m + 10 n = 0 . = (a k − bk 10 )(3 − 10 )
Case (i) n ≠ 0: = (3a k + 10bk ) − (a k + 3bk ) 10
−m = a k +1 − bk +1 10
10 = which leads to contradiction as 10 is
n ∴ the statement is also true for n = k + 1

−m i.e. by induction, the statement is true for all positive integers in.
irrational but is rational.
n Uniqueness
Case (ii) m≠0: Suppose there exists positive integers a n , bn , c n and d n such that

10n (3 + 10 ) n = a n + bn 10 and (3 + 10 ) n = c n + d n 10
10m + 10n = 0 , 10 = − which also leads to a
m Then, a n + bn 10 = c n + d n 10
contradiction. (a n − c n ) + (bn − d n ) 10 = 0 , where a n − c n , bn − d n ∈ Z
i.e. m=n=0. a n = c n and c n = d n (by (a)(ii))

Page 9
Suggested Solution for Pure Mathematics 2008 (I) Y.K.Ng (last update: 10/4/08)
(c) By (b), (3 + 10 ) n + (3 − 10 ) n = 2a n
∴ (3 − 10 ) n = 2a n − (3 + 10 ) n
Note that − 1 < 3 − 10 < 0
Case (i) n is odd:
− 1 < (3 − 10 ) n < 0
− 1 < 2a n − (3 + 10 ) n < 0
2a n < (3 + 10 ) n < 2a n − 1
∴ the greatest integer less than (3 + 10 ) n
= 2a n which is an even number, as a n ∈ Z+.
Case (ii) n is even:
0 < (3 − 10 ) n < 1
0 < 2a n − (3 + 10 ) n < 1
2a n − 1 < (3 + 10 ) n < 2a n
∴ the greatest integer less than (3 + 10 ) n
= 2a n − 1 which is an odd number, as a n ∈ Z+.

--- End of Solutions ---

Page 10