Topic 3: Algebra

Dr J Frost ([email protected])
www.drfrostmaths.com
All Maths Challenge and Olympiad problems are ©
UK Mathematics Trust (www.ukmt.org.uk)

Last modified: 16th June 2020

ζ
Topic 3 – Algebra

Part 1: Recursive Expressions and

Expansion Identities
Recursive Expressions
Sometimes values are defined in terms of themselves.

What is the value of the following?

  1
1+
1
1+
1
1+
1 +…
 If we
let the value of the above expression be , then
Multiplying both sides by x and rearranging, we get
?
Then . This is the golden ratio. Nice!
Recursive Expressions
 Question:

Without explicitly calculating or , find .

 Squaringboth sides, we get and .

Subtracting the equations (always a sensible thing we
when have two squares!), we?get
So , so
Using helpful expansion identities

In solving algebraic problems, it’s often immensely useful to find

expansions that involve the terms we know and that we’re trying
to find. For example:

 
would clearly be useful if we knew and , and wanted to find .

We used this earlier in the Number Theory module when our

equation involved , and and we wanted to factorise.
Using helpful expansion identities
 Question: We earlier found equations and
and that . Now find .

 Wecan find by adding our two equations, giving

Now look at the expansion .
Then , so ?
Using helpful expansion identities

 [SMC] Four positive integers , , and are such that . What is the
value of ?

A: 73 B: 75 C: 77

D: 79 E: 81

 
The expansion we want to use is
So
So and must have the values 1, 2, 4 and 66 (in any order). The
sum is 73.
Factorising Strategies
 
Factorise

 We might think to get these terms, we’d have something

like . Either we’d need to add an , or we’d have to add an
and before halving it.

 The expression is clearly symmetrical (in the sense we can

interchange and have the same expression), so let’s try

This gives

 Thus our factorisation is

Factorising Strategies
 
Factorise

 Notice that there’s a symmetry here between and , but not with . We’ll
have this same symmetry/asymmetry in the factorised expression.

We might attempt to start factorising like so, which would give

us the first four terms in the expansion…

  2   2
2
( 𝑎 ++𝑐
𝑏 ) ( 𝑎 +𝑏 )
…and we can see that adding this extra term in the first bracket would
give us the remaining terms.
Factorising more difficult expressions is ultimately mostly about intelligent
guess work, just by considering how terms combine across brackets.
Factorising Strategies

Try factorising these:

3 2 2 2 2 2
 
𝑥 + 𝑥 + 𝑥 𝑦 + 𝑦 =( 𝑥 + 𝑦 )(
? 𝑥+1)

 
𝑎3 − 𝑎2 +𝑎 − 1 =(𝑎 2+1 )(𝑎
? − 1)

 
𝑎𝑏+𝑏𝑐 − 3 𝑎 − 3 𝑐 = ( 𝑎+ 𝑐 ) (?𝑏 − 3 )

 Pro Tip: You can check your factorisations by trying small values for your variables,
e.g. or 1 or -1. This doesn’t guarantee it’s correct (you might have got lucky with
the values you chose!) but at least gives you greater confidence in its validity.
Using helpful expansion identities
 [BMO1] Find all integer solutions , and for:
and

You might be tempted to try and use , but this leaves the
terms and which we’re unable to deal with.

So looking at the first equation, what terms could we

combine into some factorised expression?

Then the 1st equation becomes

?
We’ll come back to this later…
Calculating big numbers

Often algebra can be used to determine large values without

a calculator. Replace numbers with variables and manipulate.
 A quickie: What is 992?

 A quickie: What is 1013?

 What about ?
You can probably get from above we get the rows of Pascal’s
triangle separated by 0s, so ?
Calculating big numbers
 [BMO1] Find the value of

Solution:
Replacing 2007 with , we get
Expanding out the brackets (make sure you’re
good on your Binomial Expansion!), we get:

At this point you might make an ‘intelligent guess’

that the numerator factorises to , which means
the overall value is . This ?is clearly much easier to
calculate.
Making an Effective Substitution

In many cases (particularly Olympiad problems), it’s

worthwhile making a substitution that simplifies our problem.

 Question: Find real solutions to

What substitution might aid us? (look at the relationship
between the contents of the three brackets)
  and
. Then
?
Now the problem is easier to solve!

  so
So
But note that and are positive,?and so is given that . So and
must be 0, and thus and are -3 and +3.
ζ
Topic 3 – Algebra

Part 2: Simultaneous Equations & Surds

Surd Manipulation
 Question: Find the least positive integer such that:

 Solution: While we could square both sides, we’ll end up with a

mixture of and . It’s better to move the to the other side first, so
that after squaring, the n’s cancel.

?
So the least positive integer is 250,000.

General Tip: Before squaring both sides, if you only have two
occurrences of variable(s), move one onto the other side of the
equality/inequality first.
Surd Manipulation

 [SMC] Which of the following is equal to:

     
A:  B:  C: 
   
D:  E: 

Hint: Perhaps we can reorganise the contents of

the outer square root such that it’s a squared
expression?
Surd Manipulation

 [SMC] Which of the following is equal to:

 
At this point, since we have twice the product of two things, it suggests we can perhaps
get the denominator in the form , so that we can take the square root. Indeed this
works very nicely:

Now we can just rationalise the denominator to get .

Solving with Surds
 
For what do the following equality hold:

 
For negative , we’d find for example that .

More generally, we have to ensure that ,

where is the modulus ?function (i.e. makes its
argument positive if not already).
Solving with Surds
 
What about:

 
We require that .
This occurs when . We could see this by sketching:

?
1
| 

1  𝒚
=
𝟏−
𝒙
Solving with Surds
 [Source: UKMT Mentoring]

Find the values of for which:

 
If we square the LHS and RHS and simplify, we end up with
.
This is the same as . Thus .
We also require that is a real value. thus .
Thus we have .

?
Simultaneous Equations

What are the two main ways you would solve simultaneous
equation?
   

Elimination Substitution

( 1 ) − ( 2 ) :2𝑥=10
   
? ?
Simultaneous Equations
BMO questions essentially only use these two principles. You just
have to be a bit more creative.

 [BMO1] Solve the simultaneous equations where are integers:

Hint: What should we substitute, and why?

 
Use substitution. Rearranging , we get .
The reason we chose to make the subject is that we eliminate
and from when we square and subtract it.
Substituting this into :
? the factor pairs of 66
It then becomes a case of considering
(including negative ones!) and working out the corresponding , , .
 Simultaneous Equations
Sometimes we add/subtract equations to be able to get the RHS
as 0 and be able to factorise the LHS.

 [BMO1] Solve for real :

Hint: They made two of the numbers a 4 for a reason!

 
From , either or . The former contradicts and ,
so . We can make this substitution to obtain:
and
? equation into the other.
We can solve this by substituting one
This gives , or .
But we need to check these! We find because it contradicts
We can use to then find for each , and any of to get .
Simultaneous Equations
Sometimes we can exploit symmetry.
 [BMO1] Solve for real :

 Hint: Notice that

 Using the above tip, let’s do:

So or . The first case contradicts so .

If we added , then we find .
?
a) (My own approach) Substitute and into .
Via elimination we’d get . Substituting this into say (1),
we get so or . If we worked out and
for the latter, would not be satisfied, so , and , thus and .
b) We could continue to add/subtract equations to work out more relationships between
variables. ?
Simultaneous Equations
Summary of Tips
 1.You can add and subtract equations:
a. When you require integer solutions, the resulting expression
after adding/subtracting might be factorisable in the form ,
where you can then reason about possible factor pairs of .
b. When your require real solutions, factorising in the form helps,
because then we can say or .
c. The above might yield the relationship between just two
variables. In which case, you could always then substitute into
the original equations to eliminate one variable.
2. You can use elimination when you can one variable in terms of
others.
3. Spot patterns in constants on the RHS. This might give clues as to
what equations would be good to add/subtract. Add/subtract based
on what variables you want in the resulting equation, rather than
just arbitrarily.
ζ
Topic 3 – Algebra

Part 3: Inequalities
Starter
 
Solve  
𝑥 ≥? 2
 
Solve  
𝑥 ≤?− 2
Dividing or multiplying by a
negative number flips the
direction of the inequality.

    1
Solve 0 ≤ 𝑥 ≤?
2
  might think you can do and hence . You’re not allowed to multiply
You
both sides by though, because you don’t know whether it’s positive or
negative, and hence it may or may not flip the direction!
You’ll learn how to solve these in FP2, but in summary, we could just
consider where for the graph , the curve has a value at least 2.
Forming inequalities using areas and lengths
By bounding a curve between two straight lines, or vice versa, we can often form some
interesting inequalities by comparing lengths and areas.

 Question: a) Using suitable squares and circles, show that .

  The fact the appears between the two bounds

a)
suggests we have a circle between two squares.
Then let the radius of the circle be 1 (so that its
area is ).
Then the outer area is 4, giving us .

If we?try comparing the areas of the circle and

smaller square we get , which is not a tight
enough bound. Comparing the lengths (of a
quarter arc with one side of the square) instead
gives us , i.e.
Forming inequalities using areas and lengths
By bounding a curve between two straight lines, or vice versa, we can often form some
interesting inequalities by comparing lengths and areas.

 Question: b) Show also (by perhaps using other shapes) that .

b)  
The screams ‘TRIANGLE!’ or ‘HEXAGON!’
since in trigonometry, angles of 30 and 60
tend to lead to lengths involving this surd.
A triangle will lead to lead to a worse bound,
since increasing the number of sides of the
surrounding polygons makes the area
?increasingly close to a circle, so let’s try a
hexagon.

Using the area of the inner hexagon, we get

which is not good enough. But comparing
lengths again gives . Compare areas to get
the upper bound of .
Solving Inequalities using Positive Terms
 
Prove that for all real .
This looks suspiciously factorisable. How therefore might
I prove this inequality?
 

The LHS is squared and therefore must be positive.

?
Thus the above inequality clearly holds.
is known as the Trivial Inequality.
 
In general, one method of proving an inequality
is to get it into the form , i.e. where the LHS is a
sum of squares.
Using this trick in Number Theory problems
 [BMO1] Find all integer solutions , and for:
and

Earlier we found we could simplify the first

equation to:

What can we determine about and ?

Since and are positive (and square), then they
both have to be 1 to add to 2.
Then and must both be .
?
We try each of the four possibilities to get our final
solution.
Solving Inequalities using Positive Terms
Practice Questions:

2
 
Prove that  
( 𝑎 − 2𝑏
? ) ≥0

    2 2
Prove that ( 𝑥 − 1? ) ≥ 0

 
(Harder!) Prove that
 
( 𝑎 − 𝑏 )2+ ( 𝑏 − 𝑐?)2+ ( 𝑐 − 𝑎 )2 ≥ 0
Using a substitution
Sometimes making a substitution makes solving an
inequality easier to solve:
 
Prove that
 
Make the substitution: ?
Then:
 

𝑥+1≥2 √ 𝑥 ?
Types of Mean

𝑨𝑩𝑪
 

Arithmetic Mean Geometric Mean

 Themean you all know and  
love: Useful for average of percentage
? changes. e.g. If your stocks and
shares account made 10% one
year, and 20% the? next, then the
average increase is , i.e. 14.8%
Harmonic Mean (not 15%!)

The average of?rates.

AM-GM Inequality
This is a hugely helpful inequality comparing Arithmetic and
Geometric Means:

Arithmetic
≥ Geometric
 

Mean Mean
 
𝑥+ 𝑦
≥ √ 𝑥𝑦
?
2
So we can form
𝑥+ 𝑦 ≥ 2 √ 𝑥𝑦
 
an inequality ?
using any sum:

How could we easily prove the above using a technique we’ve seen before?
AM-GM Inequality
Examples:
2 2
𝑥 + 𝑦 ≥ 2? 𝑥𝑦
 

4
𝑥+ 𝑦 + 𝑧 +𝑤 ≥ 4 √? 𝑥𝑦𝑧𝑤
 

2 2 2
𝑥 𝑦 + 𝑦 𝑧 + 𝑧 𝑥 ≥ 3 ?𝑥𝑦𝑧
 

𝑥 +1? ≥ 2 √ 𝑥
 

 
( 𝑥+1 ) ( 𝑦 + 1 ) ( 𝑧+1 ) ≥ 8 √? 𝑥𝑦𝑧
AM-GM Inequality
This helps us prove certain inequalities.
 
Prove that .
 
By the AM-GM inequality:
Similarly, and
Thus ?

 
Prove that
 
By the AM-GM inequality, and .
Thus
?
Cauchy-Schwarz Inequality
The Cauchy-Schwarz Inequality is really quite awesome.
Its form is as such:

This looks rather horrible, but constructing such an

inequality is simple if done is a certain way…
 Cauchy-Schwarz Inequality
Example:
STEP 1: Start with a pair of brackets on the LHS, and a
single squared bracket on the RHS.
2 2
𝑥 + 𝑥 +1 1+ 𝑦 +1
4 
2 )≥ (
2
𝑥 +)
𝑥𝑦 +1
            
()(

STEP 2: Put some sum in the RHS.

STEP 3: For each term in the sum,

square it, then think of two terms
which multiply to give this.
 Cauchy-Schwarz Inequality
We can be slightly creative with our sum.
 
“Prove that ”

1+1 
+1    
  
𝑥
( 
+ 𝑦)
 
+ 𝑧( )≥ (  
1+1+1
)2
𝑥 𝑦 𝑧
 Bythe way, this gives us a generally handy inequality of:
and similarly .
In general,
Cauchy-Schwarz Inequality
 
Given that are positive real numbers such
that , prove that

 By the Cauchy-Schwarz Inequality:

But .
So
Since is positive we can divide ?
both sides by it, as well as by 2.
 Cauchy-Schwarz Inequality
 
Let be real numbers such that .
Prove that
 (Hint: What is as a single fraction?)

 Wehave
Multiply both sides by , we get:
?
The hard part is working out what CS inequality to form!

 By
Cauchy-Schwarz,
Thus ?
Square rooting both sides, we get the desired result.
Inequality Proofs using Geometry
Sometimes there’s non-traditional and seemingly barmy ways of
proving inequalities. The following problem is on your worksheet:

 [BMO1] Prove that for all real that

 On the worksheet I recommend using the Trivial Inequality. But the
BMO model solutions mention that one student’s solution used
Heron’s Formula for the area of a triangle.
Let’s try it!

Recall that
Where ! This se c tio n o n in equalities
Coming Soon m s w il l so on by hugely
ro b le
in geometric p ck le th e very hardest
e lp ta
expanded to h
ms.
BMO1 proble
Inequality Proofs using Geometry
 Prove that for all real that

 Notice that from to , the decrease

is . Since in the formula we have
2𝑥
 
2𝑦
 
and . This suggests we should let
the sides be and . Then in Heron’s
formula:

If we square root both sides of the

inequality, the RHS matches exactly.
2𝑧
 
Now we have to think about what
area our triangle could be less than:
Inequality Proofs using Geometry
 Prove that for all real that

 We could also use the formula:

2𝑥
 
2𝑦
 
for the area of the triangle.
We can form the inequality:

?
2𝑧
 

 So far we’ve shown that using Geometry that:

All that remains is to show that

This gives , which is true by the Trivial Inequality.
Inequality Proofs using Geometry
 We saw that one potentially useful geometrical inequality is:

Another useful one is the Triangle Inequality.

𝑏
 
Triangle Inequality
 
 
𝑎
i.e. Each side is less than the sum of
the other two sides.
𝑐
  (What would happen if they were
equal?)
Summary
There’s 3 main approaches to solving inequalities:

1. Trivial Inequality
 Put your inequality in the form
This is trivially true because the sum of squares is positive.
You may need a creative factorisation.

2. AM-GM Inequality

 The arithmetic mean is greater or equal to the geometric

mean.
e.g.
Summary
There’s 3 main approaches to solving inequalities:

3. Cauchy-Schwarz
  e.g.

However, there are a number of other inequality theorems, e.g.

Rearrangement and Chebyshev inequalities, that we won’t explore here.
You can find more info here:
http://www.artofproblemsolving.com/Resources/Papers/MildorfInequali
ties.pdf