Skip to content

Mathematical problem solving

Mathematical problem solving means making progress when the route is not stated. The required techniques may be familiar, but you must decide how to represent the information, which structure matters, and whether your result answers the original question.

A reliable cycle is:

  1. Understand: identify the target, data, conditions and domain.
  2. Represent: use a diagram, table, graph, equation or simpler equivalent form.
  3. Plan: connect the representation to a known definition or result.
  4. Carry out: write a valid chain of mathematics, keeping exact values where possible.
  5. Check and reflect: test the answer, interpret it, and consider whether every case has been found.

The cycle is not rigid. A failed attempt can reveal a better representation, so you may return from carrying out a plan to understanding the problem more precisely.

You should be able to:

  • manipulate algebraic expressions and solve basic equations;
  • interpret graphs, tables and diagrams;
  • use exact values, approximations and units correctly;
  • explain a short chain of reasoning.

Review algebraic fluency if rearranging expressions requires too much attention. Review proof and reasoning if implication, counterexamples or general arguments are unfamiliar.

Separate four parts of the question.

PartQuestion to ask
TargetWhat exactly must be found, proved or decided?
DataWhich values, equations or facts are given?
ConditionsWhat restrictions or assumptions apply?
DomainWhich values or objects are allowed?

Underline command words such as exact, positive, integer, maximum and show that. They change what counts as a complete answer.

Worked example: conditions decide the solution

Section titled “Worked example: conditions decide the solution”

A positive integer is three more than another positive integer. Their product is 4040. Find the integers.

Let the smaller integer be nn. Then the larger is n+3n+3, so

n(n+3)=40.n(n+3)=40.

Rearrange and factorise:

n2+3n40=0,(n+8)(n5)=0.\begin{aligned} n^2+3n-40&=0,\\ (n+8)(n-5)&=0. \end{aligned}

Thus n=8n=-8 or n=5n=5. The algebra has two solutions, but the domain requires a positive integer. Therefore the smaller integer is 55 and the larger is 88.

Substitution checks the result:

8=5+3,5×8=40.8=5+3,\qquad 5\times8=40.

The discarded root is not an algebraic error. It is a valid solution of the equation but not of the original problem.

Misconception: every number is useful data

Section titled “Misconception: every number is useful data”

A problem may contain redundant information. Do not force every number into a calculation. Work backwards from the target and ask which facts can connect to it.

The same information can look difficult in words and simple in another form.

  • Use a diagram for geometry, forces, motion or linked lengths.
  • Use a table for a pattern, recurrence or several cases.
  • Use a graph for intersections, signs, ranges and qualitative behaviour.
  • Use an equation for an unknown quantity constrained by a relationship.
  • Use a new variable when one expression occurs repeatedly.

Name variables with their meaning and units. Writing t=t= time in seconds is more useful than introducing an unexplained xx.

A rectangle has perimeter 28 cm28\text{ cm}. Its length is 3 cm3\text{ cm} greater than its width. Find its area.

Let the width be w cmw\text{ cm}. A labelled sketch would show dimensions

wandw+3.w\quad\text{and}\quad w+3.

The perimeter condition gives

2w+2(w+3)=28.2w+2(w+3)=28.

Hence

4w+6=28,4w+6=28,

so

w=112.w=\frac{11}{2}.

The length is

w+3=172.w+3=\frac{17}{2}.

Therefore the area is

A=112×172=1874 cm2.A=\frac{11}{2}\times\frac{17}{2} =\boxed{\frac{187}{4}\text{ cm}^2}.

A common error is to write w+(w+3)=28w+(w+3)=28. That sum is only half the perimeter. The diagram makes the two pairs of equal sides visible.

Solve

22x5(2x)+4=0.2^{2x}-5(2^x)+4=0.

The expression 2x2^x repeats, and

22x=(2x)2.2^{2x}=(2^x)^2.

Let u=2xu=2^x. Then

u25u+4=0,u^2-5u+4=0,

so

(u1)(u4)=0.(u-1)(u-4)=0.

Thus u=1u=1 or u=4u=4. Return to the original variable:

2x=1x=0,2^x=1\Rightarrow x=0,

or

2x=4x=2.2^x=4\Rightarrow x=2.

Therefore x=0 or x=2\boxed{x=0\text{ or }x=2}. The substitution did not finish the problem because the target was xx, not uu.

A simpler case can reveal a pattern, test a conjecture or remove irrelevant complexity. It does not by itself prove a general claim.

Worked example: find and justify a pattern

Section titled “Worked example: find and justify a pattern”

The first nn odd positive integers are

1,3,5,,2n1.1,3,5,\ldots,2n-1.

Find their sum.

Calculate small cases:

n1+3++(2n1)11=1221+3=4=2231+3+5=9=3241+3+5+7=16=42\begin{array}{c|c} n&1+3+\cdots +(2n-1)\\ \hline 1&1=1^2\\ 2&1+3=4=2^2\\ 3&1+3+5=9=3^2\\ 4&1+3+5+7=16=4^2 \end{array}

This suggests

1+3++(2n1)=n2.1+3+\cdots +(2n-1)=n^2.

To justify it, use the arithmetic series formula. There are nn terms, with first term 11 and last term 2n12n-1:

Sn=n2(1+(2n1))=n2(2n)=n2.\begin{aligned} S_n&=\frac n2\bigl(1+(2n-1)\bigr)\\ &=\frac n2(2n)\\ &=n^2. \end{aligned}

The examples discovered the conjecture. The algebra proves it for every positive integer nn.

Misconception: checking many cases proves the rule

Section titled “Misconception: checking many cases proves the rule”

No finite list proves a statement about infinitely many values. For example, n2n+41n^2-n+41 is prime for many small positive integers, but at n=41n=41 it equals

41241+41=412,41^2-41+41=41^2,

which is not prime. Use examples to explore, then prove or disprove the resulting claim.

When the first step is unclear, ask what would be sufficient to reach the final line. Then ask what would produce that intermediate result.

The curve

y=x2+4x,x>0,y=x^2+\frac{4}{x},\qquad x>0,

has a stationary point. Find its coordinates and determine its nature.

The target needs three things:

  1. the xx coordinate, found from dy/dx=0dy/dx=0;
  2. the yy coordinate, found by substitution;
  3. classification as a maximum or minimum.

Differentiate:

dydx=2x4x2.\frac{dy}{dx}=2x-\frac{4}{x^2}.

At a stationary point,

2x4x2=0.2x-\frac{4}{x^2}=0.

Since x>0x>0, multiplying by x2x^2 is valid:

2x34=0,2x^3-4=0,

so

x3=2x=23.x^3=2\quad\Rightarrow\quad x=\sqrt[3]{2}.

Now substitute into yy. Since x3=2x^3=2, we have 4/x=2x24/x=2x^2. Therefore

y=x2+2x2=3x2=343.y=x^2+2x^2=3x^2=3\sqrt[3]{4}.

Finally,

d2ydx2=2+8x3.\frac{d^2y}{dx^2}=2+\frac{8}{x^3}.

At x=23x=\sqrt[3]{2},

d2ydx2=2+82=6>0.\frac{d^2y}{dx^2}=2+\frac82=6>0.

Hence the stationary point is the minimum

(23,343).\boxed{\left(\sqrt[3]{2},3\sqrt[3]{4}\right)}.

Notice that solving dy/dx=0dy/dx=0 alone does not answer a request for coordinates and nature.

Useful structures include symmetry, parity, factors, repeated expressions, conservation laws and quantities that remain unchanged.

Without expanding fully, solve

(x2)2=(x+4)2.(x-2)^2=(x+4)^2.

Equal squares mean

x2=x+4x-2=x+4

or

x2=(x+4).x-2=-(x+4).

The first equation gives 2=4-2=4, which is impossible. The second gives

x2=x4,x-2=-x-4,

so

2x=2x=1.2x=-2\quad\Rightarrow\quad x=-1.

Geometrically, x2=x+4|x-2|=|x+4| asks for the point equally distant from 22 and 4-4. Their midpoint is

2+(4)2=1.\frac{2+(-4)}2=-1.

Both viewpoints give x=1\boxed{x=-1}.

Starting with the pair (3,8)(3,8), one move adds 22 to both numbers. Can repeated moves produce the pair (20,25)(20,25)?

The difference between the numbers is unchanged by every move:

(b+2)(a+2)=ba.(b+2)-(a+2)=b-a.

Initially the difference is

83=5.8-3=5.

The target pair also has difference

2520=5,25-20=5,

so this invariant does not rule it out. However, after kk moves the first number is

3+2k.3+2k.

To equal 2020, we would need

3+2k=20k=172,3+2k=20\Rightarrow k=\frac{17}{2},

which is not a whole number. Therefore the target cannot be reached.

The difference condition was necessary but not sufficient. A second invariant, the parity of each number, settles the question: 3+2k3+2k is always odd, so it can never equal 2020.

Being stuck usually means that the current representation has stopped producing useful information. Do something diagnostic:

  • restate the target in symbols;
  • list the conditions you have not yet used;
  • sketch or tabulate the situation;
  • factorise, complete the square or substitute a repeated expression;
  • solve a simpler version;
  • work backwards one step;
  • test an extreme or boundary case;
  • ask whether a theorem’s conditions are satisfied.

Do not erase a failed attempt immediately. Mark why it failed. For example, “this creates two unknowns but only one equation” is useful knowledge.

Worked example: recover from an unhelpful method

Section titled “Worked example: recover from an unhelpful method”

Find the minimum value of

x+9x,x>0.x+\frac{9}{x},\qquad x>0.

Trying values may suggest a minimum near x=3x=3, but a table cannot prove it. Rewrite the gap above 66:

x+9x6=x26x+9x=(x3)2x.\begin{aligned} x+\frac9x-6 &=\frac{x^2-6x+9}{x}\\ &=\frac{(x-3)^2}{x}. \end{aligned}

Because x>0x>0 and (x3)20(x-3)^2\geq0,

(x3)2x0.\frac{(x-3)^2}{x}\geq0.

Therefore

x+9x6.x+\frac9x\geq6.

Equality occurs when (x3)2=0(x-3)^2=0, so x=3x=3. The minimum value is 6\boxed{6}.

The condition x>0x>0 is essential. Without it, dividing by xx would not preserve a non-negative expression.

Checking is part of solving, not an optional final glance.

Substitute solutions into the original equation, especially after squaring, cancelling, taking logarithms or multiplying by an expression containing a variable.

Estimate the sign and size. If 19.82/0.2119.8^2/0.21 is entered into a calculator, an estimate

2020.2=2000\frac{20^2}{0.2}=2000

will expose an output such as 18.718.7 as an entry error.

Check units, domain and physical meaning. A probability must be in [0,1][0,1]. A negative length is invalid. A modelled answer may need an integer even if the equation gives a decimal.

Ask whether there could be another case. The symbol ±\pm, endpoints, repeated roots and periodic trigonometric solutions are common sources of missing answers.

Worked example: reject an introduced solution

Section titled “Worked example: reject an introduced solution”

Solve

x+5=x1.\sqrt{x+5}=x-1.

The left side is non-negative, so any solution must satisfy x10x-1\geq0, giving x1x\geq1. Square both sides:

x+5=(x1)2.x+5=(x-1)^2.

Hence

x+5=x22x+1,x+5=x^2-2x+1,

so

x23x4=0.x^2-3x-4=0.

Factorising gives

(x4)(x+1)=0,(x-4)(x+1)=0,

so the candidates are x=4x=4 and x=1x=-1.

Check in the original equation:

x=4:9=3=41,x=4:\quad \sqrt9=3=4-1,

but

x=1:4=22.x=-1:\quad \sqrt4=2\ne-2.

Therefore the only solution is x=4\boxed{x=4}. Squaring preserves every original solution but can introduce extra ones because a2=b2a^2=b^2 does not imply a=ba=b.

  1. A number is doubled and then 77 is added, giving 3131. Form and solve an equation.
  2. Solve 32x10(3x)+9=03^{2x}-10(3^x)+9=0.
  3. Explain why the values 2,6,12,202,6,12,20 suggest, but do not prove, that the nnth term is n(n+1)n(n+1).
  4. Find the minimum value of x+16/xx+16/x for x>0x>0 by writing its difference from 88 as a non-negative expression.
  5. Solve 2x+3=x\sqrt{2x+3}=x and check all candidates.
  6. A process adds 44 to an integer on every move. Starting from 77, can it reach 3030? Give an invariant argument.
Answers
  1. Let the number be xx. Then 2x+7=312x+7=31, so 2x=242x=24 and x=12x=12.
  2. Let u=3x>0u=3^x>0. Then u210u+9=(u1)(u9)=0u^2-10u+9=(u-1)(u-9)=0, so u=1u=1 or u=9u=9. Hence x=0x=0 or x=2x=2.
  3. The formula gives the four listed values when n=1,2,3,4n=1,2,3,4, but infinitely many later terms remain unchecked. A definition, recurrence or proof is needed to establish the general term.
  4. For x>0x>0, x+16x8=(x4)2x0.x+\frac{16}{x}-8=\frac{(x-4)^2}{x}\geq0. Therefore the minimum is 88, attained at x=4x=4.
  5. Since the square root is non-negative, x0x\geq0. Squaring gives 2x+3=x22x+3=x^2, so (x3)(x+1)=0(x-3)(x+1)=0. The candidates are 33 and 1-1, but only x=3x=3 satisfies the original equation.
  6. No. Adding 44 preserves the remainder modulo 44. Every reachable number has the form 7+4k7+4k and is congruent to 33 modulo 44, whereas 3030 is congruent to 22 modulo 44.

A compact routine for unfamiliar questions

Section titled “A compact routine for unfamiliar questions”

Before writing a long solution, note:

targetfactsconditionsrepresentation\boxed{\text{target}\quad\text{facts}\quad\text{conditions}\quad\text{representation}}

Then choose one justified next step. If it fails, record what the failure teaches and change the representation. At the end, check the original problem rather than only your final equation.

Continue with mathematical modelling to learn how assumptions turn real situations into mathematics. Use calculator and technology skills to investigate and verify results responsibly. For exam questions, choosing a method and checking answers develop this process further.