← Back to subjects
0
AQA A-level Mathematics (7357) · Numerical methods
Mini-Lesson

Numerical methods

This mini-lesson covers AQA section I — Numerical methods: locating roots by change of sign (and when that fails), fixed-point iteration xn+1 = g(xn) with staircase and cobweb diagrams, the Newton–Raphson method and its failure cases, and numerical integration.

change of sign iteration x = g(x) Newton– Raphson when the algebra runs out, iterate — carefully

Work through each screen, answer the questions as you go (a few are conceptual, most are calculations) and collect ⭐ stars. Everything here is A-level standard, drawn from the AQA 7357 subject content. Press Start when you are ready.

Numerical methods · change of sign

Locating roots by change of sign

Most equations cannot be solved exactly. But if a continuous function changes sign across an interval, it must have crossed zero somewhere inside it.

If f is continuous on [a, b] and f(a) × f(b) < 0, then f(x) = 0 has at least one root in (a, b)this is the intermediate value theorem doing the work
Worked example — our running function

Let f(x) = x³ − 2x − 5.

f(2) = 8 − 4 − 5 = −1  (negative)

f(3) = 27 − 6 − 5 = 16  (positive)

f is a polynomial, hence continuous, and the sign changes. So there is a root between 2 and 3. (It is 2.0946 to 4 d.p.)

Say the magic words. Full marks require you to state that f is continuous on the interval and that there is a change of sign. Just quoting two numbers is not enough.

Numerical methods · failure

When change of sign fails

The method is a one-way test: a sign change guarantees a root, but no sign change does not guarantee no root.

  • Repeated root: the curve touches the axis without crossing (e.g. f(x) = (x − 2)²). There is a root, but no sign change.
  • Two roots in the interval: the function dips below and comes back, so the endpoint signs agree — the two crossings cancel out of the test.
  • Discontinuity: f(x) = 1/(x − 2) changes sign between 1 and 3, but has no root at all — it has an asymptote. This is why continuity is essential.
sign change ⇒ root   (for continuous f)but: no sign change ⇏ no root

Fix: use a smaller interval, or sketch the curve first. If a question asks "explain why the method fails here", the answer is nearly always a repeated root or an asymptote.

Quick check

Why a sign change works

?f is continuous on [a, b] and f(a) × f(b) < 0. What can you conclude?
Numerical methods · iteration

Fixed-point iteration x = g(x)

Rearrange f(x) = 0 into the form x = g(x), then iterate from a starting value:

xn+1 = g(xn)a root of f is a fixed point of g — a point where the curve y = g(x) meets the line y = x
Worked example — x³ − 2x − 5 = 0

Rearrange: x³ = 2x + 5 ⇒ x = (2x + 5)1/3.

Start at x₀ = 2: x₁ = (2 × 2 + 5)1/3 = 91/3 = 2.0801 (4 d.p.).

Then x₂ = (2 × 2.0801 + 5)1/3 = 2.0924, x₃ = 2.0942, x₄ = 2.0945 — converging to the root 2.0946.

The rearrangement matters enormously. The same equation also rearranges to x = (x³ − 5)/2 — and that iteration diverges. Convergence requires |g′(x)| < 1 near the root.

Sort it

Which numerical method?

Tap a description, then tap the method it belongs to.

± Change of sign

🔁 Iteration x = g(x)

📏 Newton–Raphson

Numerical methods · diagrams

Staircase and cobweb diagrams

Iterations are visualised on a graph of y = g(x) together with the line y = x. The root is where they cross. From x₀ you go vertically to the curve, then horizontally to the line, and repeat.

  • Staircase — the path climbs (or descends) towards the root in steps, all from one side. This happens when g′(x) is positive near the root.
  • Cobweb — the path spirals around the root, alternating from one side to the other. This happens when g′(x) is negative.
  • If |g′(x)| > 1 near the root, the staircase or cobweb moves away — the iteration diverges.
Staircase (g′ > 0) y = x root Cobweb (g′ < 0) y = x
Vertical to the curve y = g(x), horizontal to the line y = x, repeat. A staircase approaches from one side; a cobweb spirals in.
Quick check

The formula

?What is the Newton–Raphson iteration formula?
Numerical methods · Newton–Raphson

The Newton–Raphson method

Newton–Raphson uses the tangent at xn and takes the point where it cuts the x-axis as the next estimate. It converges very fast — typically doubling the number of correct digits each step.

xn+1 = xn − f(xn) / f′(xn)you must differentiate f — that is what the tangent needs
Worked example — f(x) = x³ − 2x − 5, x₀ = 2

f′(x) = 3x² − 2.

f(2) = −1 and f′(2) = 12 − 2 = 10.

x₁ = 2 − (−1)/10 = 2 + 0.1 = 2.1.

f(2.1) = 9.261 − 4.2 − 5 = 0.061 and f′(2.1) = 13.23 − 2 = 11.23.

x₂ = 2.1 − 0.061/11.23 = 2.1 − 0.005432 = 2.0946 (4 d.p.).

Two steps have already reached the true root 2.09455… to 4 decimal places.

Calculate

Your turn — evaluate f(2)

1Let f(x) = x³ − 2x − 5. Calculate f(2).
Hint: f(2) = 2³ − 2(2) − 5 = 8 − 4 − 5.
Calculate

Your turn — evaluate f(3)

2For the same f(x) = x³ − 2x − 5, calculate f(3). (Together with f(2), the sign change proves there is a root between 2 and 3.)
Hint: f(3) = 3³ − 2(3) − 5 = 27 − 6 − 5. Since f(2) is negative and f(3) is positive and f is continuous, a root lies in (2, 3).
Numerical methods · failure

When Newton–Raphson fails

Speed comes at a price: the method is fragile if the starting value is poorly chosen.

  • f′(xn) = 0 — the tangent is horizontal and never meets the x-axis. The formula divides by zero and the method breaks down immediately.
  • x₀ near a turning point — f′(x₀) is tiny, so f(x₀)/f′(x₀) is enormous and the tangent flings the next estimate far away, possibly converging to a completely different root.
  • x₀ at a discontinuity or where f is not differentiable — the tangent does not exist.
the danger is dividing by a very small f′(xn)always sketch, and choose x₀ close to the root and away from turning points

In the exam: "Explain why the Newton–Raphson method fails with this starting value" is nearly always answered by "f′(x₀) = 0, so the tangent is parallel to the x-axis" — and a sketch showing the tangent earns the mark.

Quick check

When it fails

?Newton–Raphson is applied with a starting value x₀ very close to a turning point of f. What is likely to happen?
Calculate

Your turn — first Newton–Raphson step

3For f(x) = x³ − 2x − 5 we have f′(x) = 3x² − 2. Starting from x₀ = 2, find x₁.
Hint: x₁ = x₀ − f(x₀)/f′(x₀). Here f(2) = −1 and f′(2) = 3(4) − 2 = 10, so x₁ = 2 − (−1)/10 = 2 + 0.1.
Numerical methods · integration

Numerical integration

The same spirit applies to areas: when ∫f(x) dx cannot be found exactly, approximate it. The trapezium rule is the method AQA requires.

ab y dx ≈ h/2 [y₀ + 2(y₁ + … + yn−1) + yn],   h = (b − a)/nn strips, n + 1 ordinates; only the interior ones are doubled
  • Convex curve (f″ > 0, bending upwards) — chords lie above the curve, so the rule overestimates.
  • Concave curve (f″ < 0) — the rule underestimates.
  • More strips → smaller h → a better estimate. Doubling n roughly quarters the error.

Improving the estimate: the standard answer to "how could the estimate be improved?" is "increase the number of strips (decrease h)". Saying "use a better calculator" earns nothing.

Match it

Match the calculation to its value (f(x) = x³ − 2x − 5)

Tap an item on the left, then its partner on the right.

Calculation
Value
Numerical methods · comparing

Choosing and comparing methods

Three tools, three purposes. Know which is which.

  • Change of signlocates a root in an interval, but does not refine it. Slow, and it needs continuity.
  • Fixed-point iteration — refines a root, requires no derivative, but only converges if |g′(x)| < 1, and the rearrangement is a genuine choice you have to make.
  • Newton–Raphson — refines a root very fast, but needs f′(x) and a good starting value.
accuracy claim: if f(2.0945) < 0 and f(2.0947) > 0, the root is 2.0946 to 4 d.p.this "show that" step is how you justify a stated accuracy

To prove an answer is correct to n decimal places, evaluate f at the two endpoints of the rounding interval and show a change of sign. Simply saying "my iteration gave 2.0946" does not demonstrate anything.

Quick check

Staircase or cobweb?

?A convergent iteration x_{n+1} = g(x_n) produces a cobweb diagram (the estimates alternate either side of the root). What does this tell you about g′(x) near the root?
Calculate

Your turn — second Newton–Raphson step

4Continue from x₁ = 2.1 to find x₂, to 4 decimal places.
Hint: f(2.1) = 9.261 − 4.2 − 5 = 0.061 and f′(2.1) = 3(4.41) − 2 = 11.23. So x₂ = 2.1 − 0.061/11.23 = 2.1 − 0.005432.
Calculate

Your turn — fixed-point iteration

5The same equation rearranges to x = (2x + 5)1/3. Starting from x₀ = 2, find x₁ to 4 decimal places.
Hint: x₁ = (2 × 2 + 5)^(1/3) = 9^(1/3) — the cube root of 9. It converges more slowly than Newton–Raphson, but it needs no derivative.
Recap

The big ideas to know

Change of sign: f continuous and f(a)f(b) < 0 ⇒ at least one root in (a, b)

It fails when: there is a repeated root, two roots, or an asymptote in the interval

Iteration: xₙ₊₁ = g(xₙ) — converges only if |g′(x)| < 1 near the root

Diagrams: staircase if g′ > 0, cobweb if g′ < 0

Newton–Raphson: xₙ₊₁ = xₙ − f(xₙ)/f′(xₙ) — fast, but fails if f′(xₙ) = 0 or x₀ sits near a turning point

Trapezium rule: more strips ⇒ better estimate; convex curve ⇒ overestimate

Justifying accuracy: show a sign change across the rounding interval

That is AQA section I — Numerical methods. Press Finish to see your score.

🏆

Mini-lesson complete!

⭐⭐⭐

You have worked through Numerical methods for AQA A-level Mathematics. 🎉

Your stars: 0 / 0

Next: test yourself in the Evaluate stage Confidence Quiz, then lock it in with Verify.

📣 Smashed it? Share your score

Challenge a mate to beat your stars, or show a parent how you got on.

→ Back to all subjects