A Potpourri of Algebra

Khor Shi-Jie
March 27, 2012
1
Experience has taught me that there are two distinct phases in the grasping of Mathematical Olympiad.
First and foremost, one has to understand the content which is being taught during Mathematical
Olympiad lessons or through reference materials. This includes all the relevant theorems of the four
major topics in MO: Algebra, Geometry, Combinatorics and Number Theory. Achieving this rst step
should not take the student too much time as it does not take much eort to learn such knowledge that is
readily available to general students of mathematics. Of much higher priority is the second phase whereby
students internalise the usage and application of these learned knowledge. One must understand and ap-
preciate various creative methods of applying learnt theorems in order to excel in MO competitions.
These notes aim to facilitate your second preparatory step in mastering the Mathematical Olympiad.
Algebra is a major topic in Junior Section, be it in the rst round or the second round. Recently
participants have complained that questions in the SMO Junior Section are becoming more challenging
and unapproachable. This is mainly due to the fact that the problems which are appearing in recent
SMO papers are fresh and some students have no experience of dealing with such questions. Nonetheless,
a strong foundation in algebra plus the application of suitable strategies will help you solve these problems.
Note my usage of the term strategies instead of theorems. A good MO student usually thinks in
terms of strategies to tackle the problem instead of theorems used to solve the problem. I have organised
past year SMO problems as well as problems from other countries according to the strategies used to
solve the problems. Hopefully this will improve your manipulation skills in solving algebra problems.
Contents
1 Substitution 3
1.1 Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Solving Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Factorisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.5 Problem Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.6 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 Completing the Square 11
2.1 Simplication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Solving Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.3 Proving Inequalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.4 Problem Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.5 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3 Factorisation 17
3.1 Solving Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2 Number Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3 Problem Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.4 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4 Miscellaneous Techniques in Algebra 24
4.1 Discriminant and Vietas theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2 Method of dierences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.3 Method of xed ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.4 Geometric constructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.5 Problem Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
4.6 Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2
Chapter 1
Substitution
The aim of using substitution (in junior section) is to simplify an expression which is otherwise too com-
plicated or untidy to study carefully. Granted, it is possible to solve such problems without substitution,
but it will take much more eort and time in competitions.
Here are several types of questions which can be solved using careful substitution:
1.1 Evaluation
In such questions, students are often given an expression with very large numbers and they are asked to
simplify or evaluate such expressions. Take a look at SMO(J)2008 First Round P23:
Evaluate
(2020
2
20100)(20100
2
100
2
)(2000
2
+ 20100)
2010
6
10
6
This is a nasty expression to work with directly without any tools, especially during a competition. We
are going to substitute the number 2010 in the equation using a variable to simplify our calculation. The
substitution x = 2010 sticks out like a sore thumb. So let us see if it works...
[(x + 10)
2
10x][(10x)
2
100
2
][(x 10)
2
+ 10x]
x
6
10
6
It looks more manageable now and can possibly be solved algebraically from now onwards, but why stop
here? Let us substitute y = 10 to further simplify the equation...
[(x +y)
2
xy][(xy)
2
y
4
][(x y)
2
+xy]
x
6
y
6
From this expression, we derive that the numerator can be simplied into:
[(x +y)
2
xy][(xy)
2
y
4
][(x y)
2
+xy] = (x
2
+xy +y
2
)y
2
(x +y)(x y)(x
2
xy +y
2
)
= y
2
(x
3
+y
3
)(x
3
y
3
)
= y
2
(x
6
y
6
)
Bingo! By using substitution we can easily see that the expression factorises and recombines to form a
sum of cube and dierence of cube expression respectively. It cancels out the denominator and produces
the elegant answer of 100.
Heres another example which is very similar to a problem that students have encountered in the MOP
class.
3
CHAPTER 1. SUBSTITUTION 4
Given that x =
3 +

5
2
, evaluate
2a
5
5a
4
+ 2a
3
8a
2
a
2
+ 1
Before substitution, we have to modify the condition given to make it look more appetising...
x =
3 +

5
2
2x + 3 =

5 x
2
3x + 1 = 0 x
2
= 3x 1
Upon multiple substitution of x
2
into the equation, one should obtain the answer of -1.
1.2 Comparison
It is useful for one to substitute a large number or an expression when asked to compare the size of two
complicated expressions. For example:
Compare the size of A =
5678901234
6789012345
and B =
5678901235
6789012347
To solve this problem, we let x be the value 5678901234 and y be the value 6789012345. We see that
A =
x
y
while B =
x + 1
y + 2
. Taking the dierence AB, we have:
x
y

x + 1
y + 2
=
2x y
y(y + 2)
which is obviously larger than 0 since 2x > y and y > 0. Hence we have A > B.
1.3 Solving Equations
The method of substitution is often used to solve equations of higher degree, exponential equations,
trigonometric equations, logarithmic equations, etc. Here are several examples:
Solve (6x + 7)
2
(3x + 4)(x + 1) = 6
I hope that the word substitution comes across your mind automatically when you see this prob-
lem. Technically, it is possible to solve this problem by expansion and using the rational root theorem.
However, this takes plenty amount of time and we should instead try to think of methods to make each
bracket look as similar as possible. This is possible by multiplying 12 to both sides of the equation,
obtaining
(6x + 7)
2
(6x + 8)(6x + 6) = (6x + 7)
2
[(6x + 7)
2
1] = 72
Upon seeing this expression, the substitution y = (6x+7)
2
should spring into your mind, simplifying the
equation into a quadratic equation with roots 9 and -8. We nally obtain x
1
=
2
3
or x
2
=
5
3
.
Let us try a problem which involves exponents:
Solve 2
4x
+ (2
x
2)
4
34 = 0
Obviously expanding (2
x
2)
4
is not wise. Again, one should think of substitution intuitively upon
seeing exponents in an equation. The trickier question is to choose the right value to substitute.
Do you gain by using the substitution y = 2
x
? Upon substitution we obtain y
4
+ (y 2)
4
34 = 0.
Once you expand the (y 2)
4
term, you have to deal with a quartic equation and that can be quite
CHAPTER 1. SUBSTITUTION 5
troublesome. Similarly, the substitution y = 2
x
2 does not help.
What if we take the value in between, say y = 2
x
1? Interestingly, the equation (y+1)
4
+(y1)
4
34 = 0
is considerably tamer because the alternate terms in the expansion cancels out each other. After ex-
panding the terms (which is simple with the aid of binomial theorem or Pascals Triangle), we get
y
4
+ 6t
2
16 = 0, which can be solved as if it is a quadratic equation. This gives us the solutions
y =

2 and y =

2. We reject the latter solution since y must be positive. Finally, we solve that
x = log
2

2 + 1.
As a side note, the theme of (x + y)
k
+ (x y)
k
is worth remembering and occurs in other kinds of
problem too. Try nding the integer closest to (2 +

2)
6
.
Finally, let us look at this mean-looking question: Evaluate
1 +
1
1 +
1
1 +
1
1 +
To solve this, let x be the expression above. Note that x is exactly the expression at the denominator of
the fraction. We then have x = 1 +
1
x
, which can be solved easily as a quadratic equation. Upon solving,
the expression is equal to
1 +

5
2
. (In case you havent realise thats the golden ratio)
1.4 Factorisation
Substitution helps in factorisation too, provided if you choose the right expression to substitute.
Factorise 4(x + 5)(x + 6)(x + 10)(x + 12) 3x
2
Do I need to expand the entire expression? Sadly, the answer is yes. However, through substitution
we can greatly simplify the process. Simply note that 5 12 = 6 10. Upon grouping useful terms
together and expanding, we obtain
4(x + 5)(x + 6)(x + 10)(x + 12) 3x
2
= 4(x + 5)(x + 12)(x + 6)(x + 10) 3x
2
= 4(x
2
+ 16x + 60)(x
2
+ 17x + 60) 3x
2
Now, y = x
2
+ 16x + 60 looks like a smart substitution. Applying that to our equation, we obtain
LHS = 4y(y +x) 3x
2
= 4y
2
+ 4xy 3x
2
= (2y x)(2y + 3x)
= (2x
2
+ 31x + 120)(2x + 35x + 120)
= (2x + 15)(x + 8)(2x
2
+ 35x + 120)
If you are an avid reader of my blog (hcmop.wordpress.com), you probably have heard about symmetric
polynomials and elementary symmetric polynomials (ESP). Heres an application of ESP to factorise the
following polynomial:
Factorise x
3
+y
3
3xy + 1
Introduce the notations
1
and
2
to indicate the ESPs for second degree polynomials x + y and xy
CHAPTER 1. SUBSTITUTION 6
respectively. We derive:
x
3
+y
3
3xy + 1 =
3
1
3xy
2
3x
2
y 3xy + 1
=
3
1
+ 1 3
1

2
3
2
= (
1
+ 1)(
2
1

1
+ 1) 3
2
(
1
+ 1)
= (
1
+ 1)(
2
1

1
3
2
+ 1)
= (x +y + 1)(x
2
+y
2
xy x y + 1)
Do read my blog for a more detailed discussion on symmetric polynomials and ESPs.
1.5 Problem Set
1. (SMO(J)2007P35) Find the largest integer N such that both N + 496 and N + 224 are perfect
squares.
2. (AIME05P7) Let x =
4
(

5 + 1)(
4

5 + 1)(
8

5 + 1)(
16

5 + 1)
. Find (x + 1)
48
.
3. (SMO(S)2011P6) Determine the value of
2
1

2 +
4

8 + 2
+
1

2
4

8 + 2
.
4. (SMO(J)2006P25) What is the product of the real roots of the equation
x
2
+ 90x + 2027
3
=
_
x
2
+ 90x + 2055
5. (SMO(J)2006P34) Suppose that the two roots of the equation
1
x
2
10x 29
+
1
x
2
10x 45

2
x
2
10x 69
= 0
are and . Find +
6. (SMO(S)2006P10) Let a and b be positive real numbers such that
1
a

1
b

1
a +b
= 0
Find the value of (
a
b
+
b
a
)
2
7. Suppose t =
a
2
a
2
+ 2b
2
+
b
2
2a
2
+b
2
. Find the minimum value of t.
8. Given that a, x, y are real numbers that satisfy
log
2
a
x + log
2
a
y log
a
(xy)
2
2, log
a
y 1
Find the range of log
a
x
2
y.
9. (SMO(O)96P3) Let a > 1 be an integer. Find all integers x such that
(
_
a +
_
a
2
1)
x
+ (
_
a
_
a
2
1)
x
= 2a
10. Solve the following the system of equations where x, y, z are positive numbers:
x
xyz
= y
2
y
xyz+1
= z
3
z
xyz+2
= x
4
CHAPTER 1. SUBSTITUTION 7
11. Solve the equation x
4
6x
3
+ 11x
2
6x + 1 = 0
12. (USAMO78P1) Given that a, b, c, d, e are real numbers such that
a +b +c +d +e = 8
a
2
+b
2
+c
2
+d
2
+e
2
= 16
Determine the maximum value of e.
CHAPTER 1. SUBSTITUTION 8
1.6 Solutions
1. There are 2 ways to solve this problem. Firstly, we can let a
2
= N + 496 and b
2
= N + 224.
We derive that a
2
b
2
= 272 (a b)(a + b) = 2
4
17. Hence, the possible solutions for
(a b, a + b) = (1, 272), (2, 136), (4, 68), (8, 34), (16, 17). Upon solving the equations, we only have
3 sets of integer solutions (a, b) = (69, 67), (36, 32), (21, 13). The largest N occurs when a is 69.
Hence we obtain N = 4265.
Another way to solve it is to note that the dierence between two squares is an even number.
To maximise N, the dierence a b must be as small as possible. N will take maximum value
when a b is 2. Hence we can construct the equation a
2
(a 2)
2
= 272 and obtain the solution
a = 69, N = 4265.
2. Substitute k =
16

5 into the equation, we obtain

x =
4
(x
8
+ 1)(x
4
+ 1)(x
2
+ 1)(x + 1)
=
4
1 +x +x
2
+ +x
15
=
4
x
16
1
x 1
=
4(x 1)
x
16
1
=
4(
16

5 1)
5 1
=
16

5 1
So we have (x + 1)
48
= 125.
3. Let x =
4

2. We have
2
1

2 +
4

8 + 2
+
1

2
4

8 + 2
=
2
1
2 +x
3
+x
2
+
1
2 x
3
+x
2
=
2
2(x
2
+ 2)
(x
2
+ 2)
2
x
6
=
(x
2
+ 2)
2
x
6
x
2
+ 2
=
(

2 + 2)
2
2

2]

2 + 2
=
6 + 2

2 + 2
= 4

2
4. Let y = x
2
+90x+2027. We have
y
3
=

y + 28. Solving yields y = 21 or y = 12. When y = 12,

there is no solution for x. When y = 21, we have x
2
+ 90x + 2006 = 0 and hence the product of
roots is 2006.
5. Let y = x
2
10x 45. We have
1
y + 16
+
1
y

2
y 24
= 0. Clearing the denominators, we obtain
y = 6. Hence we have x
2
10x 39 = 0 which suggests that the sum of roots is 10.
CHAPTER 1. SUBSTITUTION 9
6. Let x =
a
b
. We have a = bx. Through substitution, we obtain:
1
a

1
b

1
a +b
=
1
bx

1
b

1
bx +b
=
1
x
1
1
1 +x
= 0
Solving the above equation, we obtain x =

5 1
2
. This gives us (
a
b
+
b
a
)
2
= (x +
1
x
)
2
= 5.
7. Let x = a
2
+ 2b
2
, y = 2a
2
+b
2
, we have
t =
y
x +y
3
x
+
x
x +y
3
y
=
2y x
3x
+
2x y
3y
=
2y
3x
+
2x
3y

2
3
By using AM-GM inequality or the fact that m + n 2

mn, we have
2y
3x
+
2x
3y

4
3
. Hence, the
minimum value of t is
2
3
.
Using a variable to substitute the denominator is a useful technique since this allows you to simplify
terms very easily.
8. Let m = log
a
x and n = log
a
y. We have m
2
+n
2
2(m+n) 2 and n 1. Upon completing the
square, the rst equation becomes (m 1)
2
+ (n 1)
2
4. Together with the second constraint,
it forms a semicircle with radius 2 and centre at (1, 1).
We shall employ the method of linear programming to nd the range of 2m+n. Let k = 2m+n
n = 2m+k. We construct the lines y = 2x +k such that k takes the maximum and minimum
value in the graph respectively at permitted values of (x, y) as bounded by the semicircle.
All is left for us to do is to determine the y-intercept of the two linear functions. With the aid of
geometry we can derive that k [1, 3 + 2

5].
9. Let y =
_
a +

a
2
1. We also have
1
y
=
1
_
a +

a
2
1
=
_
a

a
2
1. This gives us
y
x
+
1
y
x
= 2a y
2x
2ay
x
+ 1 = 0
CHAPTER 1. SUBSTITUTION 10
By using the quadratic formula, we obtain y
x
= (
_
a +

a
2
1)
x
= a

a
2
1. Hence the
possible values for x are 2 and -2.
10. From the rst equation, we have y = x
xyz
2
. Substitute into the second equation, we have z =
x
(xyz)(xyz+1)
6
Substituting for the nal time in the third equation, we have x
(xyz)(xyz+1)(xyz+2)
6
= x
4
.
Let a = xyz. We either have x = 1 or a(a + 1)(a + 2) = 24. For the rst case, there are two
possible solutions (1, 1, 1) and (1, 1, 1). For the second case, upon rearrangement and factorisa-
tion we obtain (a 2)(a
2
+ 5a + 12) = 0. Hence, the only solution is xyz = 2. Examining the
system of equations again, we see that we either have x = y = z or x = y = z. This gives us
the two other solutions (
3

2,
3

2,
3

2) and (
3

2,
3

2,
3

2).
11. Note that x = 0 is not a solution. Dividing the equation by x
2
throughout, we have x
2
6x +11
6
x
+
1
x
2
= 0. Use the substitution y = x+
1
x
. We have y
2
6x+9 = 0 which gives us y = 3. Hence,
x +
1
x
= 3 and x =
3

5
2
.
12. From a + b + c + d = 8 e, we get the inspiration to use the substitution a =
8 e
4
+ , b =
8 e
4
+, c =
8 e
4
+, d =
8 e
4
+, where + + + = 0. We then have
a
2
+b
2
+c
2
+d
2
= 4(
8 e
4
)
2
+ 2( + + +)(
8 e
4
) +
2
+
2
+
2
+
2
= 4(
8 e
4
)
2
+
2
+
2
+
2
+
2

(8 e)
2
4
We also have a
2
+b
2
+c
2
+d
2
= 16 e
2
and hence 16 e
2

(8 e)
2
4
. Upon solving this quadratic
inequality, we have 0 e
16
5
. Hence the maximum value of e is
16
5
.
This proof is unique because it does not utilise any inequality theorems at all. Mean substitu-
tion can be useful when the sum of several variables are given.
Chapter 2
Completing the Square
Completing the square is another useful technique in simplifying expressions for closer inspection. It is
commonly applied in SMO problems when a high degree polynomial is given. This techniques is usually
used in simplication and nding extremal values.
This techniques is not restricted to expansion of squares only. Sometimes, one has to complete the
cube or expressions with higher degree. Some notable formulas include:
1. (a +b)
2
= a
2
+ 2ab +b
2
2. (a +b +c)
2
= a
2
+b
2
+c
2
+ 2ab + 2bc + 2ca
3. (a +b)
3
= a
3
+ 3a
2
b + 3ab
2
+b
3
2.1 Simplication
It is often useful to complete the squares in problems which involves variables of higher degree or expres-
sions which involves surds. For example,
Simplify
_
x 1 + 2

x 2 +
_
x 1 2

x 2
The key in this problem is to realise the fact that x 1 is equal to x 2 + 1. The original expres-
sion can be written as:
_
x 1 + 2

x 2 +
_
x 1 2

x 2 =
_
(x 2) + 2

x 2 + 1 +
_
(x 2) 2

x 2 + 1
=
_
(

x 2 + 1)
2
+
_
(

x 2 1)
2
=

x 2 + 1 +[

x 2 1[
Hence the orginal expression is equal to 2 when 2 x < 3 and 2

x 2 when x 3
2.2 Solving Equations
We know that quadratic equations can be solved using the technique of completing the squares. In fact,
the quadratic formula which produces the roots of a quadratic equation is derived using the technique of
completing the squares.
The technique of completing the squares is very useful to check if an equation has any solutions at
all. For example,
Prove that x
4
+ 3x
2
+ 2x + 6 = 0 has no real roots
To prove this, we have to complete the square in the following form:
x
4
+ 3x
2
+ 3x + 6 = (x
2
+ 1)
2
+ (x + 1)
2
+ 4 = 0
Since each term in the expression is always strictly greater than zero, there is no solution to the equation.
11
CHAPTER 2. COMPLETING THE SQUARE 12
2.3 Proving Inequalities
Probably the most useful application of completing the squares is in proving inequalities. In junior
section, students are not expected to have learned advanced inequalities such as AM-GM inequality
or Cauchy-Schwarz inequality. Students have to rely on basic algebraic manipulation skills in order to
prove inequalities in competitions. The technique of completing the squares is very handy since all squares
must be greater or equal to zero. Let us take a look at the following problem proposed by Titu Andreescu:
Let a, b, c be real numbers. Prove that the numbers a b
2
, b c
2
, c d
2
, d a
2
cannot be all larger
than
1
4
.
The solution comes intuitively if you are used to solving inequalities using the method of completing
the squares. Let us suppose that it is possible for all for expressions to be larger than
1
4
at the same
time, i.e.
a b
2
>
1
4
, b c
2
>
1
4
, c d
2
>
1
4
, d a
2
>
1
4
By adding the four expressions above together, we obtain
a +b +c +d (a
2
+b
2
+c
2
+d
2
) < 1
Moving all terms to the right and completing the squares, we have
(
1
2
a)
2
+ (
1
2
b)
2
+ (
1
2
c)
2
+ (
1
2
d)
2
< 0
which is obviously a contradiction.
Students in junior section should take note of the two fundamental ways to prove inequalities, which
are the method of dierence and the method of division respectively. To prove that A B, one can
attempt to prove that
1. AB 0 or
2.
A
B
1.
2.4 Problem Set
1. (SMO(S)2011 P3 First Round) Let x be a real number. If a = 2011x + 9997, b = 2011x + 9998,
c = 2011x + 9999, nd the value of a
2
+b
2
+c
2
ab bc ac.
2. (SMO(S)2011 P5 First Round) Suppose x, y are real numbers such that
1
x

1
2y
=
1
2x +y
. Find
the value of
x
2
y
2
+
y
2
x
2
.
3. (SMO(J)2011 P19 First Round) Let a, b, c, d be real numbers such that
a
2
+b
2
+ 2a 4b + 4 = 0
c
2
+d
2
4c + 4d + 4 = 0
Let m and M be the minimum and maximum value of (ac)
2
+(b d)
2
respectively. Find mM.
4. Suppose (x z)
2
4(x y)(y z) = 0. Prove that y is the mean of x and z.
5. Suppose x, y, z are distinct real numbers. Prove that (
1
y z
)
2
+ (
1
z x
)
2
+ (
1
x y
)
2
= (
1
y z
+
1
z x
+
1
x y
)
2
.
6. Suppose a, b, c, d are positive real numbers that satisfy a
4
+ b
4
+ c
4
+ d
4
= 4abcd. Prove that
a = b = c = d.
CHAPTER 2. COMPLETING THE SQUARE 13
7. (T. Andreescu) Find all real solutions to the system of equations
x +y =

4z 1
y +z =

4x 1
z +x =
_
4y 1
8. (IMO Longlist 1970 P37) Solve the set of simultaneous equations
v
2
+w
2
+x
2
+y
2
= 6 2z
v
2
+w
2
+x
2
+z
2
= 6 2y
v
2
+w
2
+y
2
+z
2
= 6 2x
v
2
+x
2
+y
2
+z
2
= 6 2w
w
2
+x
2
+y
2
+z
2
= 6 2v
9. (T. Andreescu) Find all real triplets (x, y, z) that satisfy x
4
+y
4
+z
4
4xyz = 1
10. Solve the system of equations
x
3
9(y
2
3y + 3) = 0
y
3
9(z
2
3z + 3) = 0
z
3
9(x
2
3x + 3) = 0
11. (SMO(J)2011 Second Round P1) Suppose a, b, c, d > 0 and x =

a
2
+b
2
, y =

c
2
+d
2
. Prove that
xy ac +bd.
12. Suppose x, y, z are real numbers that satisfy xy +yz +xz = 1. Prove that x
2
+ 5y
2
+ 8z
2
4
13. Suppose a, b, c are positive real numbers. Prove that for any real numbers x, y, z, we have
x
2
+y
2
+z
2
2

abc
(a +b)(b +c)(c +a)
(
_
a +b
c
xy +
_
b +c
a
yz +
_
c +a
b
zx)
CHAPTER 2. COMPLETING THE SQUARE 14
2.5 Solutions
1. Note that a
2
+ b
2
+ c
2
ab bc ca =
1
2
[(a b)
2
+ (b c)
2
+ (c a)
2
]. By plugging in all the
expressions in the original question, we obtain the answer 3.
2. We note that
x
2
y
2
+
y
2
x
2
= (
y
x

x
y
)
2
+ 2. From the condition given in the question, we have
1
x

1
2y
=
1
2x +y

2x +y
x

2x +y
2y
= 1

y
x

x
y
=
1
2
Hence
x
2
y
2
+
y
2
x
2
=
9
4
3. By completing the squares for the two equations given, we obtain
(a + 1)
2
+ (b 2)
2
= 1
(c 2)
2
+ (d + 2)
2
= 4
which are equations of circles with radius 1 at (-1, 2) and radius 2 at (2,-2) respectively. The
quantity (a c)
2
+ (b d)
2
represents the distance between a point in the rst circle and another
point in the second circle. Since the distance between the centres of the circles is 5 (by Pythagoras
Theorem), the maximum and minimum value of (a c)
2
+(b d)
2
are 8 and 2 respectively. Hence
the solution is 16.
4. By expanding the entire expression, we have
x
2
2xz +z
2
4xy + 4y
2
+ 4xz 4yz = 0
Upon rearrangement, we obtain
(x +z)
2
4(x +z)y + 4y
2
= 0
By completing the square, we have
[(x +z) 2y]
2
= 0
It follows that y =
x +z
2
.
5. From the fact that (x y) + (y z) + (z x) = 0, we have
1
(x y)(y z)
+
1
(y z)(z x)
+
1
(z x)(x y)
= 0. Hence,
(
1
y z
)
2
+ (
1
z x
)
2
+ (
1
x y
)
2
= (
1
y z
+
1
z x
)
2
2
1
y z

1
z x
+ (
1
x y
)
2
= (
1
y z
+
1
z x
)
2
+ 2(
1
y z
+
1
z x
)
1
x y
+ (
1
x y
)
2
= (
1
y z
+
1
z x
+
1
x y
)
2
6. We shall attempt to move all terms to left hand side, complete the squares and use the property
that no square is negative to solve this question. Since a
4
+ b
4
+ c
4
+ d
4
= 4abcd, we have
2a
4
+ 2b
4
+ 2c
4
+ 2d
4
= 8abcd. Moving all the terms to the left, we complete the squares in
the following manner:
2a
4
+ 2b
4
+ 2c
4
+ 2d
4
8abcd =(a
2
b
2
)
2
+ 2a
2
b
2
+ (b
2
c
2
)
2
+ 2b
2
c
2
+ (c
2
d
2
)
2
+ 2c
2
d
2
+ (d
2
a
2
) + 2a
2
d
2
8abcd
=(a
2
b
2
)
2
+ (b
2
c
2
)
2
+ (c
2
d
2
)
2
+ (d
2
a
2
)
2
+ 2(ab cd)
2
+ 2(bc da)
2
=0
CHAPTER 2. COMPLETING THE SQUARE 15
This gives us the following system of equations:
a
2
b
2
= 0
b
2
c
2
= 0
c
2
d
2
= 0
d
2
a
2
= 0
ab cd = 0
bc da = 0
Since a, b, c, d are positive real numbers we must have a = b = c = d.
7. We start by adding the three equations together and try to complete the squares. By taking the
sum and moving all terms to the left hand side of the equation, we have
2x + 2y + 2z

4x 1
_
4y 1

4z 1 = 0
Next, we divide the equation by 2 so that the coecient of the variables in the surds become 1.
x +y +z
_
x
1
4

_
y
1
4

_
z
1
4
= 0
Finally, we note that x
1
4

_
x
1
4
+
1
4
= (
_
x
1
4

1
2
)
2
. By completing the squares, we have
(
_
x
1
4

1
2
)
2
+ (
_
y
1
4

1
2
)
2
+ (
_
z
1
4

1
2
)
2
= 0
Since no square is negative, it follows that x = y = z =
1
2
is the only set of solution.
8. Without loss of generality, we shall discuss this problem in three dierent cases:
(a) z is not equal to any of the four other variables
(b) y = z and z is not equal to the three other variables
(c) All variables are equal
For the rst case, we nd the dierence between the rst equation and the other four equations.
We have
y
2
z
2
= 2y 2z y = 2 z
x
2
z
2
= 2x 2z x = 2 z
w
2
z
2
= 2w 2 w = 2 z
v
2
z
2
= 2v 2z v = 2 z
which implies that v = w = x = y = 2 z. Substitute back into the rst equation, we have
4(2 z)
2
= 6 2z which gives us the solutions (1,1,1,1,1) and (-1/2,-1/2,-1/2,-1/2,5/2).
For the second case, we also have the following relationships:
x
2
z
2
= 2x 2z x = 2 z
w
2
z
2
= 2w 2 w = 2 z
v
2
z
2
= 2v 2z v = 2 z
Substitute these quantities back into the rst equation , we have 3(2z)
2
+z
2
= 62z which gives
us the solutions (1,1,1,1,1) and (1/2,1/2,1/2,3/2,3/2).
Finally when all of the variables are the same, we can add all the equations and complete the
square, hence obtaining (2z +
1
2
)
2
=
125
4
which gives us the solutions (1,1,1,1,1) and (-3/2,-3/2,-
3/2,-3/2,-3/2). Hence, the nal solutions are (1,1,1,1,1), (-3/2,-3/2,-3/2,-3/2,-3/2) and permuta-
tions of (-1/2,-1/2,-1/2,-1/2,5/2) and (1/2,1/2,1/2,3/2,3/2).
P/S: Included this question initally because I thought this question can only be solved through
completing the square. Theres no need for that actually.
CHAPTER 2. COMPLETING THE SQUARE 16
9. By completing the square for the rst 2 terms, we obtain:
(x
2
y
2
) + 2x
2
y
2
+z
4
4xyz = 1
By subtracting 2z
2
from the expression, we can complete the square another time:
(x
2
y
2
)
2
+ (z
2
1)
2
+ 2x
2
y
2
+ 2z
2
4xyz = 0
Completing the square for one last time, we obtain:
(x
2
y
2
)
2
+ (z
2
1)
2
+ 2(xy z)
2
= 0
Hence, the solutions are (1, 1, 1), (1, 1, 1), (1, 1, 1).
10. We can complete the cube and rewrite the system of equations as
(y 3)
3
= y
3
x
3
(z 3)
3
= z
3
y
3
(x 3)
3
= x
3
z
3
Upon addition we obtain (x 3)
3
+ (y 3)
3
+ (z 3)
3
= 0. Without loss of generality, we assume
that x 3. From the equation z
3
9(x
2
3x + 3) = 0, we note that z
3
27 = 9x(x 3) and
hence z 3. Similarly we have y 3. This implies that the only solution to the equation is
x = y = z = 3.
11. We shall prove that x
2
y
2
(ab +cd)
2
since all numbers are non-negative. We have
x
2
y
2
= (a
2
+b
2
)(c
2
+d
2
)
= a
2
c
2
+b
2
d
2
+a
2
d
2
+b
2
c
2
a
2
c
2
+b
2
d
2
+ 2abcd
= (ab +cd)
2
due to the fact that a
2
d
2
+b
2
c
2
2abcd = (ad bc)
2
0.
We can also prove this question using properties of discriminant. Consider the quadratic equation
(a
2
+b
2
)x
2
+2(ac+bd)x+c
2
+d
2
= 0. Upon completing the squares, we obtain (ax+c)
2
+(bx+d)
2
= 0.
Since the left hand side of the equation is greater or equals to zero, the discriminant of the quadratic
equation must be smaller or equal to zero, i.e.:
4(ac +bd)
2
4(a
2
+b
2
)(c
2
+d
2
) 0 xy ac +bd
12. By shifting all terms to the left hand side of the inequality, we have:
x
2
+ 5y
2
+ 8z
2
4 = x
2
+ 5y
2
+ 8z
2
+ 4(xy +yz +zx)
= (x + 2y + 2z)
2
+ (y 2z)
2
0
13. Shifting all terms to the left, we have
x
2
+y
2
+z
2
2

abc
(a +b)(b +c)(c +a)
(
_
a +b
c
xy +
_
b +c
a
yz +
_
c +a
b
zx)
=
_
b
b +c
x
2
+
a
a +c
y
2
2

ab
(b +c)(c +a)
xy
_
+
_
c
c +a
y
2
+
b
a +b
z
2
2

bc
(c +a)(a +b)
yz
_
+
_
c
b +c
x
2
+
a
a +b
z
2
2
_
ca
(b +c)(a +b)
xz
_
=ab
_
x
_
a(b +c

y
_
b(c +a)
_
2
+bc
_
y
_
b(c +a

z
_
c(a +b)
_
2
+ca
_
z
_
c(a +b

x
_
a(b +c)
_
2
0
Hence the original inequality stands.
Chapter 3
Factorisation
Factorisation is a useful technique that is widely employed in MO. There are many things that we can
do with a factorised form of an expression. Here are some of the factorisation formula that we often
encounter:
1. ab +ac +bd +cd = (a +d)(b +c)
2. ax
2
( +)x + = a(x )(x )
3. a
2
b
2
= (a +b)(a c)
4. a
3
+b
3
= (a +b)(a
2
ab +b
2
5. a
3
b
3
= ((a b)(a
2
+ab +b
2
)
You may nd the following factorisation formulas helpful too:
1. a
3
+b
3
+c
3
3abc =
1
2
(a +b +c)[(a b)
2
+ (b c)
2
+ (c a)
2
].
2. a
4
+ 4b
4
= (a
2
+ 2ab + 2b
2
)(a
2
2ab + 2b
2
) (Sophie Germain Identity)
3. a
k
b
k
= (a b)(a
k1
+a
k2
b +a
k3
b
2
+ +b
k1
) for positive integer k
4. a
k
+b
k
= (a +b)(a
k1
a
k2
b +a
k3
b
2
+b
k
1) for odd integer k
However, the factorisation of expressions in most olympiad problems are not obvious and require tech-
niques beyond these common formulas. These techniques include:
1. Substitution
2. Coecient determination method
3. Factor theorem and rational root theorem
I have already highlighted the use of substitution to factorize expression in the rst chapter. On the other
hand, the coecient determination method is often employed when the factors have more than two terms.
For example, to factorise 2x
2
= 7xy 4y
2
3x+6y 2, we expand the expression (2xy +a)(x+4y +b)
and compare the coecients to solve for a and b.
Another very powerful tool which is often employed to factorise symmetric or cyclic polynomial is factor
theorem. The factor theorem simply states that if P(a) = 0 for some polynomial P(x), then (x a) is a
factor of the polynomial. Let us use this theorem to factorise a
3
(b c) +b
3
(c a) +c
3
(a b). By taking
a as the principle variable and evaluating P(b), we have:
P(b) = b
3
(b c) +b
3
(c b) +c
3
(b b) = 0
Hence (ab) is a factor of the expression. Similarly, (b c) and (c a) are also factors of the expression.
Now we try to let a = b c and see what we obtain:
P(b c) = (b c)
3
(b c) +b
3
[c (b c)] +c
3
[(b c) b] = 0
17
CHAPTER 3. FACTORISATION 18
This suggests that (a + b + c) is a factor too. It seems like the expression is in the form of A(a + b +
c)(a b)(b c)(c a) where A is an unknown constant. Upon verifying with the initial expression, we
obtain A = 1 and hence a
3
(b c) +b
3
(c a) +c
3
(a b) = (a +b +c)(a b)(b c)(c a).
It is often useful to check values such as a = b c to see if (a + b + c) is a factor of the expres-
sion using factor theorem (most of the time, it is). In addition, if the expression is cyclic, note that the
factors of the cyclic expression will be cyclic too. So if (ab) is a factor of the cyclic expression, naturally
(b c) and (c a) are also factors of the expression.
Finally, the rational root theorem is a tool that enables one to nd the factors of single variable polyno-
mials. The rational root theorem states that if x =
p
q
is a root of a polynomial P(x) = a
n
x
n
+ +a
0
and
(p, q) = 1, then we must have p[a
0
and q[a
n
. For example, suppose we want to factorise x
3
2x
2
2x3.
By the rational root theorem, the only possible rational roots are 1, 1, 3, 3. Upon substituting back
into the equation, we realise that x = 3 is a root of the polynomial and hence we can factorise the
polynomial into (x 3)(x
2
+x + 1).
3.1 Solving Equations
Well, I guess you have realised that this is the application that almost all techniques in algebra have in
common. The usefulness of factorisation technique is epitomised in solving equations of higher degree. In
Secondary 2, students are taught to use the method of factorisation to solve simple quadratic equations.
Using rational root theorem, we can also attempt to solve equations of higher degree by guessing the roots.
Actually, it is also possible to solve cubic equations just by using technique of factorisation without
the use of rational root theorem.
Solve the equation x
3
+px +q = 0.
Recall the factorisation a
3
+ b
3
+ c
3
3abc =
1
2
(a + b + c)[(a b)
2
+ (b c)
2
+ (c a)
2
]. We shall
use the coecient determining method to factorise the above equation into this form. The term x
3
+ q
will represent the a
3
+b
3
+c
3
portion while the term px will represent the term 3abc. We have:
a
3
+b
3
= q
ab =
p
3
We take the cube of the second equation and obtain the following system:
a
3
+b
3
= q
ab =
p
3
27
Now we can solve for a and b since a
3
and b
3
are now the roots of the quadratic equation z
2
qz
p
3
27
= 0.
We then use the quadratic formula to calculate the value of a
3
and b
3
. We nally derive that:
a =
3

q
2
+
_
q
2
4
+
p
3
27
, b =
3

q
2

_
q
2
4
+
p
3
27
We just successfully rewrote the original equation in the form of x
3
+ a
3
+ b
3
3xab = 0. Factorising
yields
1
2
(x + a + b)((x a)
2
+ (a b)
2
+ (b x)
2
) = 0. Hence to obtain x, we just have to solve the
following linear and quadratic equation:
(x +a +b) = 0 ((x a)
2
+ (a b)
2
+ (b x)
2
) = 0
Solving yields:
x
1
= a b, x
2,3
=
a +b
2

(a b)

3
2
i
CHAPTER 3. FACTORISATION 19
3.2 Number Theory
Factorisation plays an important role in number theory problems, especially in the discussion of prime
numbers and composite numbers. There were many past SMO questions on number theory which made
use of factorisation. This was a question which appeared in SMO(J) 2nd Round in 2008.
Determine all primes p such that 5
p
+ 4p
4
is a perfect square.
To solve this problem, we let 5
p
+4p
4
= n
2
for some positive integer n. We have 5
p
= (n2p
2
)(n+2p
2
).
Since all variables are integers, we infer that n + 2p
2
and n 2p
2
are certain powers of 5. So suppose
n + 2p
2
= 5
s
and n 2p
2
= 5
t
such that s + t = p and 0 s < t. We can eliminate n by taking the
dierence of the two expressions, hence obtaining 4p
2
= 5
s
(5
ts
1). Obviously if s > 0, p must be 5
and indeed the expression is a perfect square. If s = 0 and t = p, we have 5
p
= 4p
2
+1. However, we can
prove that 5
p
> 4p
2
+ 1 if p > 1 using mathematical induction. Hence, the only solution to the problem
is p = 5.
3.3 Problem Set
1. Factorise the following expressions:
(a) x
4
+x
2
+ 1
(b) x
10
+x
5
+ 1
(c) x
5
+x + 1
(d) x
9
+x
4
x 1
2. Factorise the following expression:
(a) (a b)
3
+ (b c)
3
+ (c a)
3
(b) (a + 2b 3c)
3
+ (b + 2c 3a)
3
+ (c + 2a 3b)
3
(c) (a +b +c)
3
a
3
b
3
c
3
(d) ab(a
2
b
2
) +bc(b
2
c
2
) +ca(c
2
a
2
)
(e) a
2
3b
2
8c
2
+ 2ab + 2bc + 14ca
3. Given that x
8
+x
7
+x
6
+ +x + 1 = 0, evaluate x
1998
+x
1997
+ +x
2
+x + 1.
4. Find all pairs of integers (x, y) such that x
3
+y
3
= (x +y)
2
.
5. (SMO(J) 2011 First Round P2) It is known that the roots of the equation x
5
+ 3x
4
404118x
3

12132362x
2
12132363x 2011
2
= 0 are all integers. How many distinct roots does the equation
have?
6. (SMO(S)2011 First Round P32) It is given that p is a prime number such that x
3
+y
3
3xy = p1
for some positive integers x and y. Determine the largest possible value of p.
7. (SMO(S)2006 First Round P29) Let a, b be two integers. Suppose x
2
x1 is a factor of polynomial
ax
5
+bx
4
+ 1. Find the value of a.
8. (SMO(S)2006 First Round P30) How many ordered pairs of integers (x, y) satisfy the equation
x

y +y

x +

2006xy

2006x

2006y 2006 = 0?
9. (SMO(O)2006 First Round P9) Suppose f is a function satisfying f(x + x
1
) = x
6
+ x
6
for all
x ,= 0. Determine the value of f(3)
10. (SMO(O)2008 First Round P24) Let f(x) = x
3
+ 3x + 1 where x is a real number. Given that the
inverse function of f exists and is given by
f
1
(x) =
_
x a +

x
2
bx +c
2
_
1/3
+
_
x a

x
2
bx +c
2
_
1/3
where a, b, c are positive constants, nd the value of a + 10b + 100c.
CHAPTER 3. FACTORISATION 20
11. Let r be a real number such that
3

r +
1
3

r
= 3. Determine the value of r
3
+
1
r
3
12. Prove that for every integer n > 2, the number 2
2
n
2
+ 1 is not a prime number.
13. Let a, b, c be positive real numbers satisfying (a
2
+ b
2
+ c
2
)
2
> 2(a
4
+ b
4
+ c
4
). Prove that a, b, c
must be the three sides of the same triangle.
14. (IMO 1969) Show that for any positive integers n, there exist innitely many a such that the number
n
4
+a is not prime.
CHAPTER 3. FACTORISATION 21
3.4 Solutions
1. (a) x
4
+x
2
+ 1 = x
4
+ 2x
2
+ 1 x
2
= (x
2
+ 1)
2
x
2
= (x
2
x + 1)(x
2
+x + 1)
(b) x
10
+x
5
+ 1 =
x
15
1
x
5
1
=
(x
3
1)(x
12
+x
9
+x
6
+x
3
+ 1)
(x 1)(x
4
+x
3
+x
2
+x + 1)
=
(x
2
+x + 1)(x
12
+x
9
+x
6
+x
3
+ 1)
x
4
+x
3
+x
2
+x + 1
= (x
2
+x + 1)(x
8
x
7
+x
5
x
4
+x
3
x + 1)
(c) x
5
+x + 1 = x
5
x
2
+x
2
+x + 1
= x
2
(x
3
1) +x
2
+x + 1
= x
2
(x 1)(x
2
+x + 1) + (x
2
+x + 1)
= (x
3
x
2
+ 1)(x
2
+x + 1)
(d) x
9
+x
4
x 1 = x
9
+x
5
+x
4
x
5
x 1
= (x
4
1)(x
5
+x + 1)
= (x 1)(x + 1)(x
2
+ 1)(x
3
x
2
+ 1)(x
2
+x + 1)
Note: For part (b), (c), (d), we can use complex numbers to show that x
2
+x +1 is a factor of the
expression. Let us use part (b) to illustrate this. Suppose
1
and
2
are the complex roots of the
equation x
3
1 = 0. Since (x 1)(x
2
+x + 1) = 0, we have
2
1
+
1
+ 1 = 0 and
2
2
+
2
+ 1 = 0.
We note that
10
1
+
5
1
+ 1 =
2
1
+
1
+ 1 = 0 and
10
2
+
5
2
+ 1 =
2
2
+
2
+ 1 = 0. By factor
theorem, we know that x
2
+x + 1 must be a factor of the expression.
2. (a) By using the property that a
3
+ b
3
+ c
3
= 3abc when a + b + c = 0, we have (a b)
3
+ (b
c)
3
+ (c a)
3
= 3(a b)(b c)(c a).
(b) By using the same property as the previous part, we have (a + 2b 3c)
3
+ (b + 2c 3a)
3
+
(c + 2a 3b)
3
= 3(a + 2b 3c)(b + 2c 3a)(c + 2a 3b).
(c) Let a be the principal variable of the function. Suppose a = b. We have
f(b) = c
3
b
3
+b
3
c
3
= 0
By factor theorem, we know that (a + b) is a factor of the expression. Since the expression
is a cyclic expression, (b + c) and (c + a) must be factors of the expressions too. Hence, the
factorised form of the expression must be A(a +b)(b +c)(c +a) where A is a constant. Upon
expansion, we solve that A = 3 and the factorised form of the expression is 3(a+b)(b+c)(c+a).
(d) Let a be the principal variable of the function. Suppse a = b. We have
f(b) = b
2
(b
2
b
2
) +bc(b
2
c
2
) bc(c
2
b
2
) = 0
Hence (a b) is a factor of the expression. Since the expression is a cyclic, (b c) and (c a)
are also the factors of the expression. Now we suppose a = b c. We have
f(b c) = (b c)[(b +c)
2
b
2
] +bc(b
2
c
2
) +c(b c)[c
2
(b +c)
2
] = 0
Therefore (a + b + c) is a factor of the expression. The original expression is in the form of
A(a + b + c)(a b)(b c)(c a). Upon expansion and checking the coecients, we derive
A = 1 and hence the factorised form is (a +b +c)(b a)(c b)(a c).
(e) Note that a
2
+2ab 3b
2
= (a +3b)(a 2b). We suppose that the factorised expression of the
equation is (a+3b +mc)(ab +nc) for some constants m, n that we want to determine. Upon
expansion, we obtain a
2
3b
2
8c
2
+2ab +2bc +14ca = a
2
3b
2
+mnc
2
+2ab +(3nm)bc +
(m+n)ca. We need to solve the following system of equations:
m+n = 2
3n m = 14
mn = 8
CHAPTER 3. FACTORISATION 22
Solving yields m = 2, n = 4. Hence the factorised expression is (a + 3b 2c)(a b + 4c)
3. From the condition we have x
9
+x
8
+ +x = 0. Hence, x
1998
+x
1997
+ +1 = (x
9
+x
8
+ +
x)(x
1989
+x
1980
+ +x
9
+ 1) + 1 = 1.
4. Obviously, (k, k) is a solution for all integers k. Suppose x ,= y. We have x
2
xy +y
2
= x +y.
We can write this expression as x
2
(y + 1)x + y
2
y = 0. By using the quadratic formula, we
have
x =
y + 1
_
(y + 1)
2
4(y
2
y)
2
=
y + 1
_
6y 3y
2
+ 1
2
Hence 6y 3y
2
+ 1 0 and since y is an integer, y can only be 1 or 2. Hence, the ve possible
solutions are (k, k), (1, 1), (2, 1), (1, 2), (2, 2)
5. By using the rational root theorem, we know that the roots are either 1, 1, 2011, 2011, 2011
2
or
2011
2
. Since the sum of roots is 3 while the product of roots is 2011
2
, the only possible set of
solutions is (1, 1, 1, 2011, 2011). Hence there are 3 distinct solutions.
6. Introduce the notations
1
and
2
to indicate the ESPs for second degree polynomials x + y and
xy respectively. We derive:
x
3
+y
3
3xy + 1 =
3
1
3xy
2
3x
2
y 3xy + 1
=
3
1
+ 1 3
1

2
3
2
= (
1
+ 1)(
2
1

1
+ 1) 3
2
(
1
+ 1)
= (
1
+ 1)(
2
1

1
3
2
+ 1)
= (x +y + 1)(x
2
+y
2
xy x y + 1)
= p
Since x +y +1 > 1, we must have x +y +1 = p and x
2
+y
2
xy x y +1 = 1. From the second
equation, we have x
2
+ y
2
xy x y = 0. By using the result in problem 4, we know that the
only solutions are (1, 1), (2, 1), (1, 2), (2, 2). Hence, the largest prime p is 5 when x = y = 2.
7. Suppose the factorised form of the expression is (x
2
x 1)(ax
3
+cx
2
+dx 1). Upon expansion,
we have ax
5
+bx
4
+1 = ax
5
+(c a)x
4
+(dc a)x
3
(1+d+c)x
2
+(1d)x+1. By comparing
coecients, we obtain the following system of equations:
d c a = 0
1 +d +c = 0
1 d = 0
Which gives us the solution d = 1, c = 2, a = 3. Hence the solution is a = 3.
8. Let k =

2006,
1
=

x +

y and
2
=

xy. We have
x

y +y

x +
_
2006xy

2006x
_
2006y 2006 =
1

2
+k
2
k
1
k
2
=
1
(
2
k) +k(
2
k)
= (
1
+k)(
2
k)
= (

x +

y +

2006)(

xy

2006) = 0
Hence we must have xy = 2006. There are 8 possible solutions:
(1, 2006), (2, 1003), (17, 118), (34, 59), (59, 34), (118, 17), (1003, 2), (2006, 1)
9. f(x +x
1
) = x
6
+x
6
= (x
2
+x

2)(x
4
2 +x
4
)
= [(x +x
1
)
2
2][(x
2
+x
2
)
2
4]
= [(x +x
1
)
2
2][(x +x
1
)
2
2]
2
4
Hence, f(x) = (x
2
2)[(x
2
2)
2
4]. We have f(3) = 322.
CHAPTER 3. FACTORISATION 23
10. Let y = x
3
+3x+1. By shifting all terms to the left we have x
3
+3x+1 y = 0. We want to write
this in the form of x
3
+ a
3
+ b
3
3abx = 0 so that we can factorise the expression. By comparing
terms, we have:
a
3
+b
3
= 1 y
ab = 1
From the second equation, we have a
3
b
3
= 1. Hence, a
3
and b
3
are the roots of the quadratic
equation t
2
(1 y)t 1 = 0. By using the quadratic formula, we have:
a =
3

1 y +
_
y
2
2y + 5
2
b =
3

1 y
_
y
2
2y + 5
2
Upon factorisation of x
3
+ a
3
+ b
3
3abx we obtain (x + a + b)(x
2
+ a
2
+ b
2
ax bx ab) = 0.
From the rst factor, we obtain x = a b =
3

y 1 +
_
y
2
2y + 5
2
+
3

y 1
_
y
2
2y + 5
2
.
This suggests that f
1
(x) =
3
_
x 1 +

x
2
2x + 5
2
+
3
_
x 1

x
2
2x + 5
2
, and hence a =
1, b = 2, c = 5. The value of 100c + 10b +a is 521.
11. We use the property a
3
+ b
3
+ c
3
= 3abc when a + b + c = 0 to solve this problem. From the
condition, we have
1
3

r
+
3

r 3 = 0. Hence, r +
1
r
27 = 3(
3

r)(
1
3

r
) = 9 and r +
1
r
= 18.
Similarly, r
3
+
1
r
3
18
3
= 3(r
3
)(
1
r
3
)(18) = 54. Therefore r
3
+
1
r
3
= 18
3
54 = 5778.
12. Using the Sophie Germain Identity, we have 1+2
2
n
2
= (1+2
2
n2
+2
2
n1
1
)(12
2
n2
+2
2
n1
1
).
It suces to prove that 1 2
2
n2
+ 2
2
n1
1
> 1, which is obviously true since 2
2
n1
1
> 2
2
n2
for
n > 2.
13. Upon expansion and moving all the terms to the left hand side, we have
(a
2
+b
2
+c
2
)
2
2(a
4
+b
4
+c
4
) = 2a
2
b
2
+ 2b
2
c
2
+ 2c
2
a
2
a
4
b
4
c
4
= (a
2
b
2
)
2
c
4
+c
2
(2a
2
+ 2b
2
)
= (a +b)
2
(a b)
2
c
4
+c
2
[(a +b)
2
+ (a b)
2
]
= [(a +b)
2
c
2
][(a b)
2
c
2
]
= (a +b +c)(a +b c)(a b +c)(a +b +c) > 0
Since the expression is larger than 0, we either have all four factors to be larger than 0 or two of
the factors to be larger than zero. Suppose it is the latter case. Without loss of generality, let us
assume that a b + c < 0 and b + c a < 0. Taking the sum of these two inequalities, we obtain
2c < 0 which is a contradiction. Hence, all four factors of the expression must be larger than zero.
We have:
a +c > b
a +b > c
b +c > a
Since the three numbers satisfy the triangle inequality, they can form sides of a triangle.
14. We shall prove that n
4
+4k
4
is a composite number for all positive integers k. By using the Sophie-
Germain Identity, we have n
4
+ 4k
4
= (n
2
+ 2nk + 2k
2
)(n
2
2nk + k
2
). It suces to prove that
n
2
2nk + 2k
2
> 1, which is true since (n k)
2
+k
2
1 > 0. Since there is an innite amount of
choices for k, the proposition is proven.
Chapter 4
Miscellaneous Techniques in Algebra
4.1 Discriminant and Vietas theorem
The discriminant of a quadratic equation is given by the expression = b
2
4ac. It is a well-known
fact that a quadratic equation has 2 distinct real roots if > 0, two equal real roots when = 0 and
two distinct complex roots when < 0. If it is stated that a quadratic equation has two real roots, we
can deduce that 0. (Note that this condition includes the situation when the two roots are equal. I
remember that I used to regard equal roots as one real root and answered an SMO problem wrongly).
Most students are aware of the application of discriminant to determine the number of real roots. The
discriminant can also come in handy if it is stated that the roots of the quadratic equation are integers
or rational numbers. In both cases, the root of the discriminant must be a rational number i.e. the
discriminant must either be a perfect square or the square of a rational number.
Problem one in the second round of SMO(J)2006 illustrates this idea:
Find all integers (x, y) that satisfy the following equation:
x +y = x
2
xy +y
2
Rearranging yields x
2
(y + 1)x + y
2
y = 0. Since the solutions are integers, this implies that the
discriminant is a perfect square. The discriminant of this equation is = 6y 3y
2
+ 1, which is larger
than zero only when y = 0, 1, 2. It turns out that the discriminant is a perfect square for these values,
and this gives us the 6 solutions (x, y) = (0, 0), (1, 0), (0, 1), (2, 1), (1, 2), (2, 2).
On the other hand, Vietas theorem relates the elementary symmetric polynomials in terms of the roots
of a polynomial to the coecients of the variable. For a quadratic equation ax
2
+bx +c = 0, we have
x
1
+x
2
=
b
a
x
1
x
2
=
c
a
For equations of higher degree such as a
n
x
n
+a
n1
x
n1
+ +a
1
x +a
0
= 0, we have:

1
=

sym
x
i
= x
1
+x
2
+ +x
n
=
a
n1
a
n

2
=

sym
x
i
x
j
= x
1
x
2
+x
1
x
3
+ +x
2
x
3
+x
2
x
4
+ +x
n1
x
n
=
a
n2
a
n
.
.
.

n
= x
1
x
2
x
n
= (1)
n
a
0
a
n
Vietas theorem is widely applied in MO problems in various forms. It is also useful to apply Vietas
theorem if the signs of the roots are given, since the signs of the coecients are dependent on the signs
of each root.
24
CHAPTER 4. MISCELLANEOUS TECHNIQUES IN ALGEBRA 25
4.2 Method of dierences
The essence of this method is to break down an expression into dierences and eliminate terms by taking
telescoping sum. This method is often applied (but not limited to) if the expression contains fractions or
factorials.
To split a fraction into its corresponding dierence expression, we often employ the method of partial
fractions. Generally we have:
1.
k
(ax +b)(cx +d)
=
A
ax +b
+
B
cx +d
2.
k
(ax +b)
2
=
A
ax +b
+
B
(ax +b)
2
3.
k
(ax
2
+bx +c)(dx +e)
=
Ax +B
ax
2
+bx +c
+
C
dx +e
Apart from coecient determining method, theres a shortcut that allows us to deduce the constants of
the numerator of the partial fractions quickly if the denominators are linear expressions. To illustrate
this shortcut, consider the problem below.
Find a general formula for the sum
1
1 2 3
+
1
2 3 4
+ +
1
n(n + 1)(n + 2)
We want to decompose the fraction
1
n(n+1)(n+2)
into its partial fractions. To nd the constant above
the n term, we can ignore the n part in the original fraction and substitute n = 0 into the fraction. This
gives us the value
1
2
, which will be the constant term of the partial fraction. Similarly, we ignore the
expression n + 1 and substitute n = 1 into the fraction, deriving -1 as the constant term of the partial
fraction with denominator n + 1. Finally we ignore the expression n + 2 and substitute n = 2 into the
fraction, obtaining
1
2
as the constant term of the nal partial fraction with denominator n +2. We have:
1
n(n + 1)(n + 2)
=
1
2n

1
n + 1
+
1
2(n + 1)
By telescoping sum, we have:
1
1 2 3
+
1
2 3 4
+ +
1
n(n + 1)(n + 2)
=
1
2

1
2
+
1
4
+
1
2(n + 1)

1
n
+
1
2(n + 2)
=
1
4

2n
2
+ 4n + 1
2n(n + 1)(n + 2)
Heres a problem which involves factorials.
(SMO(S)1999P24) Evaluate 1 1! + 2 2! + 3 3! + + 9 9!.
Note that n n! = (n + 1)! n!. Using this fact, we have
1 1! + 2 2! + 3 3! + + 9 9! = 10! 1 = 362879.
4.3 Method of xed ratios
This technique is applicable to questions that has several equal quantities such as a = b = c = d. It is
usually more useful for problems with a xed ratio or quantities in exponential form. Given a condition
a
b
=
c
d
=
e
f
, we can let this quantity equate to a constant k and derive a = bk, c = dk, e = fk which
may be useful in solving the problem.
On the other hand, to manage exponential relationships like a
x
= b
y
= c
z
, we can equate this quan-
tity to another constant k and take logarithm to obtain the relationship x = log
a
k, y = log
b
y and
z = log
c
z.
CHAPTER 4. MISCELLANEOUS TECHNIQUES IN ALGEBRA 26
Consider this problem:
(AIME85P7) Suppose a, b, c, d are positive integers that satisfy
a
5
= b
4
c
3
= d
2
c a = 19
Evaluate d b.
From the rst condition, we can let a
5
= b
4
= e
2
0 for some integer e. Also, we can let c
3
= d
2
= f
6
for some integer f. From the third condition, we have f
2
e
4
= 19 (f e
2
)(f + e
2
) = 19. Since
19 is a prime, we must have f e
2
= 1, f + e
2
= 19. Solving yields e = 3 and f = 10. Hence,
d b = 10
3
3
5
= 757.
4.4 Geometric constructions
This class of problems are very tricky because without the inspiration to use diagrams to simplify the
problem it can be very tedious to solve the problem. One must be very sensitive to algebraic expressions
which appear like a geometric formula such as the Pythagorean Theorem, cosine rule, equations of circles,
etc. The following problem was take from SMO(J) last year:
(SMO(J)2011 First Round P14) Let a, b, c be positive real numbers such that
a
2
+ab +b
2
= 25
b
2
+bc +c
2
= 49
c
2
+ca +a
2
= 64
Find (a +b +c)
2
.
We need to know the value of ab + bc + ca in order to evaluate the quantity (a + b + c)
2
. It is ob-
viously impossible to calculate this value directly from the system of equations. However, we notice that
the right hand side of the equations are all perfect squares. Also, recall the cosine rule which states that
a
2
= b
2
+c
2
2bc cos A for triangle ABC. It looks somewhat similar to the system of equations, provided
that the angle in the cosine rule expression is 120 degrees. Of course, it is impossible to construct a
triangle such that each angle is 120 degrees. We can, however, construct three lines from a point such
that the angle between each line is 120 degrees. Consider the triangle below:
We let [AD[ = a, [BD[ = b, [CD[ = c. Using cosine rule, we derive that [AB[ = 5, [BC[ = 7, [CA[ = 8.
We still need to nd out the value of ab + bc + ca. Consider the sine formula for area of triangle which
states that S
ABC
=
1
2
bc sin A. We note that the area of the large triangle is equal to the sum of areas of
the three small triangles. By equating these two quantities using Herons formula and sine formula for
area, we have:
1
2
(ab +bc +ca) sin 120

=
_
10(5)(2)(3)
ab +bc +ca = 40
Finally, we have (a +b +c)
2
= a
2
+b
2
+c
2
+ 2(ab +bc +ca) =
1
2
(25 + 49 + 64 + 3 40) = 129.
CHAPTER 4. MISCELLANEOUS TECHNIQUES IN ALGEBRA 27
4.5 Problem Set
1. (SMO(S)2009 First Round P3) If two real numbers a and b are randomly chosen from the interval
(0, 1), nd the probability that the equation x
2

ax +b = 0 has real roots.

2. (SMO(S)2006 First Round P35) Let p be an integer such that both roots of the equation 5x
2

5px + (66p 1) = 0 are positive integers. Find the value of p.

3. (AIME2008P7) Let r, s, and t be the three roots of the equation 8x
3
+ 1001x + 2008 = 0. Find
(r +s)
3
+ (s +t)
3
+ (t +r)
3
.
4. (AIME2011P15) Let P(x) = x
2
3x9. A real number x is chosen at random from the interval 5
x 15. The probability that
_
P(x)| =
_
P(x|) is equal to

a +

b +

c d
e
, where a, b, c, d
and e are positive integers and none of a, b, c, d and e are positive integers. Find a +b +c +d +e.
5. Compute 1!(1
2
+ 1 + 1) + 2!(2
2
+ 2 + 1) + 3!(3
2
+ 3 + 1) + + 9!(9
2
+ 9 + 1).
6. Compute 1 2 + 2 3 + 3 4 + +n (n + 1).
7. (SMO(J)2006 First Round P9) Find the value of
1
3 + 1
+
2
3
2
+ 1
+
4
3
4
+ 1
+ +
2
2006
3
2
2006
+ 1
.
8. (Ukraine) Prove the inequality
1

1 +

3
+
1

5 +

7
+ +
1

9997 +

9999
> 24.
9. (Canada) Given
a
1
b
1
=
a
2
b
2
=
a
3
b
3
and p
1
, p
2
, p
3
are not all equal to zero. Prove that for all positive
integers n, we have (
a
1
b
1
)
n
=
p
1
a
n
1
+p
2
a
n
2
+p
3
a
n
3
p
1
b
n
1
+p
2
b
n
2
+p
3
b
n
3
.
10. Solve the following systems of equations:
(a)
x
2
+y
2
+z
2
x +y
=
14
3
x
2
+y
2
+z
2
y +z
=
14
5
x
2
+y
2
+z
2
z +x
=
7
2
(b) ax = by = cz =
1
x
+
1
y
+
1
z
(a > 0, b > 0, c > 0)
11. Given that n is a natural number, ax
2
n
+1
= by
2
n
+1
= cz
2
n
+1
, and
1
x
+
1
y
+
1
z
= 1. Prove that
2
n
+1
_
ax
2
n
+by
2
n
+cz
2
n
=
2
n
+1

a +
2
n
+1

b +
2
n
+1

c.
12. Suppose x
1
, x
2
, , x
100
are 100 positive numbers that satisfy the following conditions:
(a) x
2
1
+x
2
2
+ +x
2
100
> 10000
(b) x
1
+x
2
+ +x
100
300
Prove that it is possible to nd 3 numbers such that the sum of these 3 numbers is larger than 100.
13. (AIME2006P15) Given that x, y and z are real numbers that satisfy:
x =
_
y
2

1
16
+
_
z
2

1
16
y =
_
z
2

1
25
+
_
x
2

1
25
z =
_
x
2

1
36
+
_
y
2

1
36
and that x +y +z =
m

n
, where m and n are positive integers and n is not divisible by the square
of any prime, nd m+n.
CHAPTER 4. MISCELLANEOUS TECHNIQUES IN ALGEBRA 28
4.6 Solutions
1. The discriminant of the quadratic equation is a 4b. Since the equation has real roots, we must
have 0 and hence a 4b. The triangular region below indicates the possible values of a and b
such that the inequality holds within the domain a, b (0, 1). Since the area of the triangle is
1
8
,
the probability when the equation has real roots is
1
8
1 =
1
8
-
X
X
X
X
X
X
X
X
X
X
XX
0 1
a
6
0
1
4
1
b
2. Suppose the two roots of the equation are m and n. We have m + n = p. Since m, n are positive
integers, p must be a positive integer too. Since the quadratic equation has integer coecients and
integer solutions, the discriminant of the equation must be a perfect square. Suppose = k
2
for
some integer k. We have:
(5p)
2
4(5)(66p1) = k
2
(5p132)
2
k
2
= 17404 (5p132+k)(5p132k) = 419229
Since the expressions 5p132+k and 5p132k have the same parity and since 17404 is divisible
by 4, both expressions must be even numbers. We must have:
_
5p 132 +k = 458
5p 132 k = 38
or
_
5p 132 +k = 38
5p 132 k = 458
The rst set of equation gives us p = 76 while the second set of equations give us a negative value
of p. Hence the only solution for p is 76.
3. By using Vietas theorem, we have the following relationships:
r +s +t = 0
rs +st +tr =
1001
8
rst = 251
Since r +s +t = 0, we have:
(r +s)
3
+ (s +t)
3
+ (t +r)
3
= 3(r +s)(s +t)(t +r)
= 3(rs
2
+r
2
s +st
2
+s
2
t +tr
2
+t
2
r + 2rst)
= 3[(r +s +t)(rs +st +tr) rst]
= 3(251)
= 753
4. This is a sinister problem that looks complicated but its existence merely serves the purpose of
wasting your time in the competition. Needless to say this brutal problem did serve its purpose
well. Firstly, note that P(x|) must be a perfect square since
_
P(x)| is a positive integer. By
checking all the integers in between 5 and 15 inclusively, we realise that P(x) is a positive integer
when x = 5, 6, 13.
CHAPTER 4. MISCELLANEOUS TECHNIQUES IN ALGEBRA 29
When x [5, 6), we have x
2
3x9| = 1. For the equation to hold, the value of x
2
3x9 must
be less than 4. By using quadratic equation to solve for x
2
3x 9 < 4, we obtain x <
3 +

61
2
.
Since this number is smaller than 6, we obtain our rst valid interval x [5,
3 +

61
2
).
When x [6, 7), we have x
2
3x9| = 3. For the equation to hold, the value of x
2
3x9 must
be less than 16. By using quadratic equation to solve for x
2
3x9 < 16, we obtain x <
3 +

109
2
.
Since this number is smaller than 7, we obtain our second valid interval x [6,
3 +

109
2
).
When x [13, 14), we have x
2
3x9| = 11. For the equation to hold, the value of x
2
3x9 must
be less than 144. By using quadratic equation to solve for x
2
3x9 < 144, we obtain x <
3 +

621
2
.
Since this number is smaller than 14, we obtain our rst valid interval x [13,
3 +

621
2
).
To compute the probability of the equation being valid, we have:
P(
_
P(x)| =
_
P(x|)) =
(
3 +

61
2
5) + (
3 +

109
2
6) + (
3 +

621
2
13)
10
=

61 +

109 +

621 39
20
The nal solution is 850.
5. Generally, we have
k!(k
2
+k + 1) = k![(k + 1)
2
k]
= (k + 1)(k + 1)! k k!
By telescoping sum, we have 1!(1
2
+1 +1) +2!(2
2
+2 +1) +3!(3
2
+3 +1) + +9!(9
2
+9 +1) =
10 10! 1 = 36287999.
6. Generally, we have n(n+1) =
1
3
(n(n+1)(n+2) (n1)(n)(n+1)). By telescoping sum, we have
1 2 + 2 3 + +n (n + 1) =
1
3
n(n + 1)(n + 2)
7. Generally, we have
2
k
3
2
k
+ 1
=
2
k
3
2
k
1

2
k
3
2
k
1
+
2
k
3
2
k
+ 1
=
2
k
3
2
k
1

2
k+1
3
2
k+1
1
By telescoping sum,
1
3 + 1
+
2
3
2
+ 1
+
4
3
4
+ 1
+ +
2
2006
3
2
2006
+ 1
=
1
2

2
2007
3
2
2007
1
8. We observe that

1
2
+

3
2
+

5
2
+ +

10001

9999
2
> 48
In addition, we note that
1

1 +

3
+
1

5 +

7
+ +
1

9997 +

9999
>
1

3 +

5
+
1

7 +

9
+
+
1

9999 +

10001
. After rationalising the denominators, we must have
1

1 +

3
+
1

5 +

7
+ +
1

9997 +

9999
> 24
, which proves our proposition.
CHAPTER 4. MISCELLANEOUS TECHNIQUES IN ALGEBRA 30
9. Let
a
1
b
1
=
a
2
b
2
=
a
3
b
3
= k where k is not equal to zero. We obtain three equations a
1
= kb
1
, a
2
=
kb
2
, a
3
= kb
3
. Now we have:
p
1
a
n
1
+p
2
a
n
2
+p
3
a
n
3
p
1
b
n
1
+p
2
b
n
2
+p
3
b
n
3
=
p
1
(kb
1
)
n
+p
2
(kb
2
)
n
+p
3
(kb
3
)
n
p
1
b
n
1
+p
2
b
n
2
+p
3
b
n
3
= k
n
= (
a
1
b
1
)
n
which proves the proposition.
10. (a) Let k = x
2
+ y
2
+ z
2
, a = x + y, b = y + z, c = z + x. From the system of equations, we have
14
3
a =
14
5
b =
14
4
c = k, which suggests that a : b : c = 3 : 5 : 4. We let a = 3, b = 5, c = 4
for some constant . Solving for x, y, z we obtain x = , y = 2, z = 3. Substitute back
into the rst equation, we have
14
2
3
=
14
3
which gives us = 1. Hence, our solution is
(x, y, z) = (1, 2, 3).
(b) Suppose ax = by = cz = k for some constant k. We have x =
k
a
, y =
k
b
, z =
k
c
. We also have
1
x
+
1
y
+
1
z
= k which upon substitution gives us
a +b +c
k
= k and k =

a +b +c. Hence,
our solution is x =
1
a

a +b +c, y =
1
b

a +b +c, z =
1
c

a +b +c.
11. Let ax
2
n
+1
= by
2
n
+1
= cz
2
n
+1
= k where k is some arbitrary constant. We have ax
2n
=
k
x
, by
2n
=
k
y
, cz
2n
=
k
z
and
2
n
+1
_
ax
2n
+by
2n
+cz
2n
=
2
n
+1
_
k(
1
x
+
1
y
+
1
z
) =
2
n
+1

k. On the other hand,

2
n
+1

a =
2
n
+1

k
x
,
2
n
+1

b =
2
n
+1

k
y
,
2
n
+1

c =
2
n
+1

k
z
. The sum of these three expressions,
2
n
+1

a +
2
n
+1

b +
2
n
+1

c = (
1
x
+
1
y
+
1
z
)
2
n
+1

k =
2
n
+1

k. Hence the equality in the question holds.

12. We will prove this problem by contradiction. Let us suppose that it is not possible to nd three
numbers such that the sum of these three numbers is more than 100. Without loss of generality,
let us assume that x
1
x
2
x
3
x
100
. Our assumption suggests that x
1
+ x
2
+ x
3
< 100.
We suppose that x
1
.x
2
, x
3
, x
100
are the sides of 100 squares. Place all these squares in a straight
line side by side from x
1
to x
100
. Consider the diagram below
We superpose these squares onto a grid of 3 big squares, each with length 100. Since x
1
+ x
2
+
+ x
100
< 300, the squares will t within the three squares. Our assumption also suggests that
x
1
, x
2
, x
3
will t into the rst grid. Now, we divide part of the rst grid into three rectangles as
shown in the diagram above, each with the length of 100 and height of x
1
, x
2
, x
3
respectively. We
can move all the other squares from the second big square into the rectangle with height x
2
and
CHAPTER 4. MISCELLANEOUS TECHNIQUES IN ALGEBRA 31
we can move all the squares from the third big square into the rectangle with the height x
3
. This
suggests that the sum of the areas of these squares is less than 100
2
as they are able to t within
the rst big square. This contradicts with the condition x
2
1
+ x
2
2
+ x
2
100
> 10000 and hence the
assumption is false. This would mean that there must be three numbers (in this case x
1
, x
2
, x
3
)
such that their sum is larger than 100.
13. On hindsight, this looks like another AIME problem which serves to bore contestants by forcing
them to square each equations, get rid of radicals and solve the simultaneous quadratic equations.
However, theres an elegant solution which makes use of the following geometric construction.
Consider the following triangle with sides AB = x, AC = y, BC = z
From the system of equations and by Pythagorean theorem, it seems that the height of the triangle,
AD, BE, CF are
1
6
.
1
5
,
1
4
respectively. There are two methods that we can use to solve for x, y, z at
this juncture:
(a) By considering the area of the triangle, we obtain the relationship
1
4
x =
1
5
y =
1
6
z. By using
method of xed ratios, we can let x = 4k, y = 5k, z = 6k for some arbitrary constant k. Upon
substitution into the original equation, we can solve for k and obtain the value of x, y and z.
(b) Similarly, by using the relationships established by the area of triangle, we have y =
5
4
x and
z =
6
4
x. We can use Herons formula to set up the following equation:
1
8
=
_
(
15
8
x)(
7
8
x)(
5
8
x)(
3
8
x)
which enables us to solve for x easily and derive the value of y and z.
I prefer the second approach. Solving yields x =
8
15

7
. Hence x+y +z =
8
15

7
+
10
15

7
+
12
15

7
=
2

7
. The nal answer is 9.
The above procedure is sucient for AIME since AIME only requires candidates to give their
nal answer. However, one requires the following substantiations to prove the uniqueness of the
solution:
(a) To prove that this is the only possible set of altitudes and the only possible triangle that can
be constructed, we observe that the equation x =
_
y
2
h
2
+

z
2
k
2
is strictly increasing
for LHS and strictly decreasing for RHS. Hence, there can only be one possible set of altitude
for triangles with the sides x, y, z based on the system of equations.
(b) To show that there exist a triangle with side lengths x, y, z, we observe that

x
2
k
2
<
x,
_
y
2
k
2
< y,

z
2
k
2
< z. From the system of equations, we have x < y + z, y <
z +x, z < x +y. Hence the three variables satisfy the triangle inequality.