← AS & A Level Mathematics

AS & A Level Mathematics · Pure Mathematics 2 & 3 · Chapter 6

Numerical solutions of equations

🎯 Learning objectives

  • Understand why numerical methods are needed when algebraic solutions are unavailable.
  • Locate approximately a root of an equation by sketching a suitable pair of intersecting graphs.
  • Use the change-of-sign rule for a continuous function to show that a root lies within an interval [a, b].
  • Identify when the change-of-sign rule fails or requires caution (asymptotes, discontinuities, even multiplicity).

Book exercises: Exercise 6A (Finding a starting point), Exercise 6B (Improving your solution), Exercise 6C (Applications to other areas of mathematics), End-of-chapter review exercise 6, and Cross-topic review exercise 2.

Why solve equations numerically?

  • Algebraic limits: You are familiar with solving linear and quadratic equations algebraically using factoring or the quadratic formula. However, higher-degree equations (such as quintics x⁵ + x - 1 = 0) or transcendental equations mixing algebraic, exponential, or trigonometric terms (such as ex = 3x + 2 or x = tan 2x) have no exact general algebraic solution.
  • Numerical methods: These calculate approximate solutions to any required level of precision by generating a sequence of values that converge to the root α.
  • The starting point: Every numerical method requires an initial approximation x₁ reasonably close to the root to ensure convergence.

Method 1: The graphical approach

  • Splitting the equation: To locate the roots of f(x) = 0, rearrange the equation into the form g(x) = h(x), where y = g(x) and y = h(x) are standard functions with familiar graphs.
  • Intersection points: The x-coordinates of points where the two curves intersect are precisely the real roots of f(x) = 0.
  • Number of real roots: The number of intersection points gives the total number of real roots in the given domain.
122468xyα ≈ 1.38y = x³y = 4 − x
Figure 6.1: Intersecting graphs y = x³ and y = 4 - x reveal a single real root α ≈ 1.38 between x = 1 and x = 2.

Worked Example 6.1: Locating a root graphically

By sketching graphs of y = x³ and y = 4 - x, show that the equation x³ + x - 4 = 0 has a root α between 1 and 2.

  1. Step 1: Rearrange the equation to equate two simple functions: x³ = 4 - x.
  2. Step 2: Sketch y = x³ (cubic through origin with stationary inflection) and y = 4 - x (straight line with gradient −1 and intercepts (0, 4) and (4, 0)).
  3. Step 3: At x = 1, x³ = 1 and 4 - x = 3, so the line is above the curve (4 - x > x³).
  4. Step 4: At x = 2, x³ = 8 and 4 - x = 2, so the curve is above the line (x³ > 4 - x).
  5. Step 5: Conclude: the curves intersect exactly once for x > 0, and the x-coordinate of this intersection satisfies 1 < α < 2. Therefore x³ + x - 4 = 0 has a root α between 1 and 2. ✔

Method 2: The change-of-sign rule

  • Core theorem: If a function f(x) is continuous on the closed interval [a, b], and f(a) and f(b) have opposite signs (i.e. f(a) × f(b) < 0), then there exists at least one value α between a and b such that f(α) = 0.
  • Why it works: A continuous curve has no breaks, jumps, or holes. To travel from below the x-axis to above the x-axis, the curve must cross the x-axis at least once.
  • Exam requirement: Always state explicitly: (1) evaluate f(a) and f(b), (2) state the signs ("f(a) < 0 and f(b) > 0"), and (3) write the conclusion: "Change of sign of continuous function indicates presence of a root between a and b".
0.51−11xyf(0) = −1 < 0f(1) = 1 > 0Root α ≈ 0.755y = x⁵ + x − 1 (continuous)
Figure 6.2: Sign change for continuous f(x) = x⁵ + x - 1: f(0) = -1 < 0 and f(1) = 1 > 0 guarantees a root α ≈ 0.755.

Worked Example 6.2: The change-of-sign test

Show by calculation that the equation f(x) = x⁵ + x - 1 = 0 has a root α between 0 and 1.

  1. Step 1: Check continuity: f(x) = x⁵ + x - 1 is a polynomial, so it is continuous for all real x.
  2. Step 2: Evaluate at lower limit: f(0) = 0⁵ + 0 - 1 = -1 < 0.
  3. Step 3: Evaluate at upper limit: f(1) = 1⁵ + 1 - 1 = 1 > 0.
  4. Step 4: Conclude: because f(x) is continuous and changes sign between x = 0 and x = 1, there must be at least one root α such that 0 < α < 1 and f(α) = 0. ✔

When the change-of-sign test fails or needs care

  • Discontinuities & vertical asymptotes: If f(x) has an asymptote in [a, b], f(a) and f(b) can have opposite signs without any root existing! For instance, f(x) = (1)/(x - 1) has f(0.5) = -2 < 0 and f(1.5) = 2 > 0, but there is an asymptote at x = 1, not a root. Similarly, f(x) = tan x has a sign change across x = π/2.
  • Repeated roots (even multiplicity): When a curve touches the x-axis without crossing it (e.g. f(x) = (x - 2)² or f(x) = |tan x|), f(x) ≥ 0 on both sides. Hence f(1.9) > 0 and f(2.1) > 0: there is NO change of sign, even though x = 2 is a genuine root!
  • Multiple roots in interval: If an interval contains an even number of roots (e.g. 2 roots), the curve crosses the axis and crosses back, leaving f(a) and f(b) with the same sign. Keep intervals sufficiently narrow.
12−22xyAsymptote at x = 1 (No root)f(a) < 0f(b) > 0123123xyTouching at x = 2 (Root, no sign change)f(a) > 0f(b) > 0Root at x = 2
Figure 6.3: Pitfalls of the sign-change test: (left) asymptote causes a sign change without a root; (right) touching curve has a root with no sign change.

Worked Example 6.3: Trigonometric equation with sign change

(a) By sketching suitable graphs, show that cos x = 2x - 1 has only one root for 0 ≤ x ≤ π/2. (b) Verify by calculation that this root lies between 0.8 and 0.9.

  1. Step 1: Sketch y = cos x (starts at (0, 1) and reaches (π/2, 0)) and y = 2x - 1 (line through (0, -1) and (0.5, 0)).
  2. Step 2: For 0 ≤ x ≤ π/2, the graphs intersect exactly once, so the equation has only one root.
  3. Step 3: Define f(x) = cos x - 2x + 1 (working in radians!).
  4. Step 4: Evaluate at x = 0.8: f(0.8) = cos(0.8) - 2(0.8) + 1 = 0.69671 - 1.6 + 1 = +0.09671 > 0.
  5. Step 5: Evaluate at x = 0.9: f(0.9) = cos(0.9) - 2(0.9) + 1 = 0.62161 - 1.8 + 1 = -0.17839 < 0.
  6. Step 6: Conclude: since f(x) is continuous and changes sign, the root lies between 0.8 and 0.9. ✔
0.511.5−10.51xyα ≈ 0.89y = cos xy = 2x − 1
Figure 6.4: Single intersection of y = cos x and y = 2x - 1 confirmed on [0.8, 0.9].

✏️ Try it

  1. Show by calculation that the equation x³ + 2x - 5 = 0 has a root α between 1 and 2.
  2. By sketching the graphs of y = ex and y = 4 - x, determine the number of real roots of ex + x - 4 = 0.
  3. Explain why the change-of-sign test on f(x) = (1)/(x - 2) on the interval [1, 3] does not indicate a root.
Answers

1. Step 1: Let f(x) = x³ + 2x - 5, which is a continuous polynomial. Step 2: f(1) = 1 + 2 - 5 = -2 < 0. Step 3: f(2) = 8 + 4 - 5 = 7 > 0. Step 4: Change of sign indicates a root α lies between 1 and 2.

2. Step 1: Rearrange as ex = 4 - x. Step 2: Sketch exponential curve y = ex (strictly increasing, y-intercept (0, 1)) and straight line y = 4 - x (gradient −1, y-intercept (0, 4)). Step 3: The line and curve intersect at exactly one point, so there is 1 real root.

3. Step 1: f(1) = 1/(1 - 2) = -1 < 0 and f(3) = 1/(3 - 2) = 1 > 0. Step 2: Although there is a change of sign, f(x) has a vertical asymptote at x = 2 and is discontinuous on [1, 3]. Step 3: The change of sign is due to the infinite discontinuity, not an axis crossing; there is no root.

Next: 6.2 Iteration & rearrangements →

Rearranging f(x) = 0 into x = F(x)

  • The principle: Any root α of f(x) = 0 satisfies f(α) = 0. By algebraic rearrangement, we isolate one x on the left-hand side: x = F(x). At the root, α = F(α).
  • Fixed point: On a graph, the solution to x = F(x) is the point of intersection between the straight line y = x and the curve y = F(x).
  • Recurrence formula: Using subscript notation, we define the iterative formula:
    xₙ₊₁ = F(xₙ)
    Given an initial approximation x₁, we compute x₂ = F(x₁), then x₃ = F(x₂), and so on.
123123xy(α, α) with α ≈ 1.379y = xy = ∛(4 − x)
Figure 6.5: Fixed-point intersection: solving x³ + x - 4 = 0 via x = ∛(4 - x) where y = x meets y = F(x) at (α, α).

How the iterative sequence works

  • Mapping steps: Start at (x₁, x₁) on y = x. Move vertically to (x₁, F(x₁)) on the curve y = F(x). The resulting y-coordinate is x₂. Then move horizontally to (x₂, x₂) on y = x. Repeat vertically to find x₃.
  • Multiple rearrangements: An equation can be rearranged into x = F(x) in several ways. For example, x³ + 5x - 7 = 0 can be rewritten as:
    • x = (7 - x³)/(5) (isolating x from the linear term 5x)
    • x = ∛(7 - 5x) (isolating x from the cubic term x³)
    • x = (7)/(x² + 5) (factoring x(x² + 5) = 7)
    Crucially, not all rearrangements converge to the root! Some diverge away from the root.
11.5211.52xyStart (x₁, x₁)(x₁, F(x₁)) = (x₁, x₂)(x₂, x₂)
Figure 6.6: Visualising successive iterates: vertical movement evaluates F(xₙ); horizontal movement transfers the value back to the line y = x.

Worked Example 6.4: Rearranging and generating iterations

The equation x² + x - 3 = 0 has a negative root α.
(a) Show that the equation can be rearranged as x = (3)/(x) - 1.
(b) Use the iterative formula xₙ₊₁ = (3)/(xₙ) - 1 with x₁ = -2.5 to find α correct to 2 decimal places. Give each iteration to 4 decimal places.

  1. Step 1: Rearrange algebraically: x² + x - 3 = 0 ⇒ x² + x = 3. Divide through by x (since x ≠ 0): x + 1 = (3)/(x) ⇒ x = (3)/(x) - 1. ✔
  2. Step 2: Set up iterative formula: xₙ₊₁ = (3)/(xₙ) - 1 with x₁ = -2.5.
  3. Step 3: Calculate x₂ = (3)/(-2.5) - 1 = -1.2 - 1 = -2.2000.
  4. Step 4: Calculate x₃ = (3)/(-2.2) - 1 = -2.3636.
  5. Step 5: Calculate x₄ = (3)/(-2.3636) - 1 = -2.2692.
  6. Step 6: Calculate x₅ = -2.3220, x₆ = -2.2920, x₇ = -2.3089, x₈ = -2.2993, x₉ = -2.3047.
  7. Step 7: Notice that x₈ = -2.30 and x₉ = -2.30 both round to -2.30 to 2 d.p. Candidate root is α ≈ -2.30. ✔

Using your calculator efficiently with the Ans key

  • Step 1: Type the initial guess x₁ (e.g. -2.5) and press =. This stores the value into the calculator's Ans memory.
  • Step 2: Type the formula using Ans in place of xₙ: e.g. 3 / Ans - 1.
  • Step 3: Press = repeatedly. Each press computes and displays the next iterate x₂, x₃, x₄, ….
  • Accuracy rule: Never round intermediate values inside the calculator; keep full internal precision and write down intermediate steps to the requested number of decimal places (typically 4 to 6 d.p.).
Equation f(x) = 0e.g. x³ + 5x − 7 = 0Isolate x = F(x)Multiple options possibleCheck |F′(α)| < 1Ensures convergence
Figure 6.7: The iterative rearrangement pipeline from initial equation to confirmed convergence.

Worked Example 6.5: Exponential equation iteration

The equation ex - 2x - 1 = 0 has a positive root β between 1 and 2.
(a) Show that this equation can be rearranged as x = ln(2x + 1).
(b) Use the iteration xₙ₊₁ = ln(2xₙ + 1) with x₁ = 1.5 to find β to 3 significant figures, recording each step to 5 significant figures.

  1. Step 1: Isolate ex: ex = 2x + 1. Take natural logarithms of both sides: ln(ex) = ln(2x + 1) ⇒ x = ln(2x + 1). ✔
  2. Step 2: Start with x₁ = 1.5000.
  3. Step 3: x₂ = ln(2(1.5) + 1) = ln(4) = 1.3863.
  4. Step 4: x₃ = ln(2(1.38629) + 1) = 1.3278.
  5. Step 5: x₄ = ln(2(1.32776) + 1) = 1.2962.
  6. Step 6: x₅ = 1.2788, x₆ = 1.2691, x₇ = 1.2636, x₈ = 1.2605.
  7. Step 7: Since x₇ ≈ 1.26 and x₈ ≈ 1.26, the root appears to be β = 1.26 (3 s.f.). ✔

✏️ Try it

  1. Show that the equation x³ - 3x - 5 = 0 can be rearranged into the form x = ∛(3x + 5).
  2. Use the iterative formula xₙ₊₁ = ∛(3xₙ + 5) with x₁ = 2 to find x₂, x₃ and x₄ correct to 4 decimal places.
  3. State an alternative rearrangement of x³ - 3x - 5 = 0 by isolating x from the linear term, and explain why it might not converge near α ≈ 2.28.
Answers

1. Step 1: Add 3x + 5 to both sides: x³ = 3x + 5. Step 2: Take the cube root: x = ∛(3x + 5). ✔

2. Step 1: x₁ = 2. Step 2: x₂ = ∛(3(2) + 5) = ∛(11) ≈ 2.2240. Step 3: x₃ = ∛(3(2.22398) + 5) = ∛(11.67194) ≈ 2.2683. Step 4: x₄ = ∛(3(2.26834) + 5) = ∛(11.80502) ≈ 2.2769.

3. Step 1: Rearrange as 3x = x³ - 5 ⇒ x = (x³ - 5)/(3). Step 2: For F(x) = (x³ - 5)/(3), the derivative is F′(x) = x². Step 3: At the root α ≈ 2.28, F′(α) ≈ (2.28)² ≈ 5.2 > 1. Since |F′(α)| > 1, this formula diverges and will fail to find the root.

Next: 6.3 Cobwebs, staircases & convergence →

Cobweb and staircase diagrams

  • Cobweb diagram (oscillating / spiral):
    • Occurs when the gradient of y = F(x) at the root is negative: F′(α) < 0.
    • Iterates alternate: if xₙ < α, then xₙ₊₁ > α, and vice versa.
    • If -1 < F′(α) < 0, the rectangle steps spiral inward towards (α, α): convergent cobweb.
    • If F′(α) < -1, the rectangle steps spiral outward away from (α, α): divergent cobweb.
  • Staircase diagram (monotonic):
    • Occurs when the gradient of y = F(x) at the root is positive: F′(α) > 0.
    • Iterates approach the root monotonically from one direction without crossing to the other side.
    • If 0 < F′(α) < 1, the steps climb or descend towards the intersection: convergent staircase.
    • If F′(α) > 1, the steps move progressively further away: divergent staircase.
0.811.21.41.60.811.21.41.6xyx₁ = 1.45Root α ≈ 1.137−1 < F′(α) < 0 (Cobweb)
Figure 6.8: Inward Cobweb: -1 < F′(α) < 0 produces alternating iterates spiralling toward the root.
11.21.41.61.811.21.41.61.8xyx₁ = 1.8Root α ≈ 1.2560 < F′(α) < 1 (Staircase)
Figure 6.9: Monotonic Staircase: 0 < F′(α) < 1 produces one-sided steps approaching the root.

The mathematical condition for convergence

  • Convergence theorem: The iteration xₙ₊₁ = F(xₙ) converges to a root α (for any initial value x₁ sufficiently close to α) if and only if: \[|F'(\alpha)| < 1\]
  • Rate of convergence:
    • If |F′(α)| ≈ 0, convergence is extremely rapid.
    • If |F′(α)| is close to 1 (e.g. 0.95), convergence is very slow.
    • If |F′(α)| > 1, the iteration diverges.
0.811.21.41.60.811.21.41.6xyx₁ = 1.10|F′(α)| > 1: Spirals OUTWARDS (Diverges)
Figure 6.10: Divergent iteration: |F′(α)| > 1 causes successive iterates to spiral outward away from the intersection.

🔬 Interactive simulation: Cobweb & Staircase Iteration Visualiser

Select a formula and adjust the starting value x₁ and number of iterations to observe live whether the iterates form an inward/outward cobweb or a staircase.

Interactive dynamic plot of y = x, y = F(x), and iteration steps.

    Worked Example 6.6: Analysing convergence analytically

    For the equation x³ + 5x - 7 = 0, two possible rearrangements are:
    (1) x = (7 - x³)/(5) and (2) x = ∛(7 - 5x).
    Given that the root α ≈ 1.1370, determine analytically which formula converges and describe the diagram pattern for each.

    1. Step 1: For formula (1), F₁(x) = (7 - x³)/(5) ⇒ F₁′(x) = -(3x²)/(5).
    2. Step 2: Evaluate at root: F₁′(1.1370) = -(3(1.1370)²)/(5) ≈ -0.7757.
    3. Step 3: Since -1 < F₁′(α) < 0, |F₁′(α)| < 1. Therefore formula (1) converges and forms an inward cobweb pattern. ✔
    4. Step 4: For formula (2), F₂(x) = (7 - 5x)1/3 ⇒ F₂′(x) = (1)/(3)(7 - 5x)−2/3(-5) = −(5)/(3(7 − 5x)2/3).
    5. Step 5: Since α³ = 7 - 5α, (7 - 5α)2/3 = (α³)2/3 = α².
    6. Step 6: Evaluate at root: F₂′(1.1370) = -(5)/(3(1.1370)²) ≈ -1.289.
    7. Step 7: Since |F₂′(α)| ≈ 1.29 > 1, formula (2) diverges and produces an outward spiralling cobweb. ✔

    ✏️ Try it

    1. For F(x) = ln(3x + 2) near root α ≈ 1.76, calculate F′(α) and state whether the iteration gives a cobweb or staircase.
    2. An equation is rearranged as x = (4)/(x² + 1). If the root is α ≈ 1.38, evaluate |F′(α)| and deduce whether the iteration converges.
    3. Explain why the iteration xₙ₊₁ = 2xₙ - 1 diverges when searching for the root α ≈ 2.5.
    Answers

    1. Step 1: F(x) = ln(3x + 2) ⇒ F′(x) = (3)/(3x + 2). Step 2: At α = 1.76, F′(1.76) = (3)/(3(1.76) + 2) = (3)/(7.28) ≈ 0.412. Step 3: Since 0 < F′(α) < 1, the sequence converges and produces a convergent staircase pattern.

    2. Step 1: F(x) = 4(x² + 1)−1 ⇒ F′(x) = -4(x² + 1)−2(2x) = -(8x)/((x² + 1)²). Step 2: At α = 1.38, α² + 1 = 1.38² + 1 ≈ 2.9044. Step 3: F′(1.38) = -(8(1.38))/((2.9044)²) = -(11.04)/(8.4355) ≈ -1.309. Step 4: Since |F′(α)| ≈ 1.31 > 1, the formula diverges.

    3. Step 1: For F(x) = 2x - 1, the derivative is F′(x) = 2x ln 2. Step 2: At α = 2.5, F′(2.5) = 22.5 ln 2 ≈ 5.657 × 0.693 ≈ 3.92. Step 3: Because F′(2.5) ≈ 3.92 > 1, each step increases the distance from α, causing divergence.

    Next: 6.4 Confirming accuracy & bounds →

    The need for formal verification

    • The consecutive iterate trap: Students often assume that because x₅ = 1.3787 and x₆ = 1.3789 agree to 2 decimal places (1.38), the root is guaranteed to be 1.38. However, if convergence is slow (|F′(α)| close to 1), consecutive iterates may differ by very little while remaining far from the true root α!
    • Cambridge rule: In A Level examinations, you must never rely purely on consecutive iterates to justify accuracy. You must use the change-of-sign test on the upper and lower bounds of the required accuracy interval.
    1.4651.471.475−0.040.04xyf(1.465) < 0f(1.475) > 0Trapped in intervalα
    Figure 6.11: Rigorous interval bounding: establishing that f(1.465) < 0 and f(1.475) > 0 traps α in (1.465, 1.475), proving α = 1.47 to 2 d.p.

    How to set up the bounds test

    • To confirm a root to d decimal places:
      • If the candidate root is A, the half-unit interval is:
        [A − 0.5 × 10⁻ᵈ, A + 0.5 × 10⁻ᵈ]
      • For 1 decimal place (A = 1.4): test 1.35 and 1.45.
      • For 2 decimal places (A = 1.47): test 1.465 and 1.475.
      • For 3 decimal places (A = 1.473): test 1.4725 and 1.4735.
    • To confirm a root to k significant figures:
      • For 3 significant figures (A = 1.26): the final significant digit is in the hundredths place (0.01), so test 1.255 and 1.265.
      • For 3 significant figures (A = 24.3): the final digit is in the tenths place (0.1), so test 24.25 and 24.35.
    • Writing the proof:
      1. Substitute both bound values into the original equation f(x).
      2. State clearly that one value is positive and the other is negative.
      3. Conclude: "Since f(x) is continuous and changes sign between L and U, the root α lies in (L, U). Therefore α = A correct to the required accuracy."
    123123xyTrue root α|x₂ − x₁| is tiny... yet x₁ is far from α! (Why bounds are needed)
    Figure 6.12: Slow convergence trap: when the gradient of F is close to 1, steps are small even when far from α.

    Worked Example 6.7: Complete root confirmation proof

    An iterative process for f(x) = x² + x - 3 = 0 produces the estimate α ≈ -2.30. Prove by calculation that α = -2.30 correct to 2 decimal places.

    1. Step 1: Identify the bounds for 2 decimal places: -2.30 ± 0.005 ⇒ [-2.305, -2.295].
    2. Step 2: Note that f(x) = x² + x - 3 is continuous on this interval.
    3. Step 3: Evaluate at lower bound: f(-2.305) = (-2.305)² + (-2.305) - 3 = 5.313025 - 2.305 - 3 = +0.008025 > 0.
    4. Step 4: Evaluate at upper bound: f(-2.295) = (-2.295)² + (-2.295) - 3 = 5.267025 - 2.295 - 3 = -0.027975 < 0.
    5. Step 5: Conclude: since f(-2.305) > 0 and f(-2.295) < 0 and f(x) is continuous, the root α lies in the interval (-2.305, -2.295).
    6. Step 6: Since every number in this interval rounds to -2.30 to 2 decimal places, α = -2.30 correct to 2 decimal places is proven. ✔

    🛠️ Change-of-Sign & Fixed-Point Iteration Calculator

    Select a preset equation and rearrangement, enter a starting value x₁ and target accuracy, and inspect both the sequence of iterates and the formal interval change-of-sign test.

      ✏️ Try it

      1. An iterative formula produces successive values x₆ = 1.3787 and x₇ = 1.3789. State the candidate root to 2 decimal places and the bounds that must be tested.
      2. Given f(x) = x³ + x - 4 = 0, evaluate f(1.375) and f(1.385) to confirm that the root is 1.38 correct to 2 decimal places.
      3. To confirm that a root is α = 0.842 correct to 3 decimal places, what two values must be substituted into f(x)?
      Answers

      1. Step 1: Candidate root is 1.38 (2 d.p.). Step 2: The bounds are 1.38 ± 0.005, which gives 1.375 and 1.385.

      2. Step 1: f(1.375) = (1.375)³ + 1.375 - 4 = 2.5996 + 1.375 - 4 = -0.0254 < 0. Step 2: f(1.385) = (1.385)³ + 1.385 - 4 = 2.6568 + 1.385 - 4 = +0.0418 > 0. Step 3: Change of sign on continuous f(x) confirms that α lies in (1.375, 1.385), so α = 1.38 (2 d.p.).

      3. Step 1: Degree of accuracy is 3 decimal places, so the half-unit error is 0.0005. Step 2: Lower bound is 0.842 - 0.0005 = 0.8415. Step 3: Upper bound is 0.842 + 0.0005 = 0.8425.

      Next: 6.5 Applications & problem solving →

      Geometric applications: Circular sectors & segments

      • Circular measure review: For a sector of radius r and angle θ radians:
        • Arc length s = rθ
        • Area of sector Asector = (1)/(2)r²θ
        • Area of triangle Atriangle = (1)/(2)r²sinθ
        • Area of segment Asegment = (1)/(2)r²(θ - sinθ)
      • Equating areas: Real-world design constraints often link sector and segment areas. Setting Asegment = Atriangle gives (1)/(2)r²(θ - sinθ) = (1)/(2)r²sinθ ⇒ θ = 2sinθ, which must be solved numerically.
      θC (centre)ABShaded triangleUnshaded segmentr = 8 cmLogo: Area(segment) = Area(triangle) ⇒ θ = 2 sin θ
      Figure 6.13: Logo design where unshaded segment area equals shaded triangle area, generating θ = 2sinθ.

      Worked Example 6.8: Sector and segment logo problem

      In a company logo ABC, AB is an arc of a circle with centre C and radius AC = 8 cm. The area of the unshaded segment is equal to the area of the shaded triangle ABC, and \angle ACB = θ radians.
      (a) Show that θ = 2sinθ.
      (b) Use the iteration θₙ₊₁ = 2sinθₙ with θ₁ = 1.85 to find θ to 3 significant figures.
      (c) Find the arc length AB.

      1. Step 1: Area of triangle = (1)/(2)r²sinθ; Area of segment = (1)/(2)r²(θ - sinθ).
      2. Step 2: Equate areas: (1)/(2)r²(θ - sinθ) = (1)/(2)r²sinθ ⇒ θ - sinθ = sinθ ⇒ θ = 2sinθ. ✔
      3. Step 3: Iterate with θ₁ = 1.85:
        • θ₂ = 2sin(1.85) = 1.9226
        • θ₃ = 2sin(1.9226) = 1.8775
        • θ₄ = 2sin(1.8775) = 1.9066
        • θ₅ = 1.8883, θ₆ = 1.9001, θ₇ = 1.8926, θ₈ = 1.8974, θ₉ = 1.8943, θ₁₀ = 1.8963, θ₁₁ = 1.8950
      4. Step 4: The values converge to θ = 1.90 (3 s.f.).
      5. Step 5: Verify bounds on f(θ) = θ - 2sinθ: f(1.895) = -0.0016 < 0 and f(1.905) = +0.0163 > 0. Change of sign confirms θ = 1.90.
      6. Step 6: Calculate arc length: s = rθ = 8 × 1.8955 = 15.2 cm (to 3 s.f.). ✔

      Calculus applications: Stationary points & optimization

      • Transcendental derivatives: Finding the maximum, minimum, or points of inflection for curves involving products of exponentials, logs, or trigonometric functions often leads to non-algebraic derivative equations:
        dy/dx = 0 ⇒ g(x) = 0
      • Parametric curves: On curves given by x = f(t), y = g(t), the chain rule gives (dy)/(dx) = (g′(t))/(f′(t)). Setting (dy)/(dx) = m often produces an equation in t that requires numerical solution.
      0.10.20.30.40.020.040.06xyMax P (dy/dx = 0)x ≈ 0.27
      Figure 6.14: Stationary point on y = x²cos 4x: (dy)/(dx) = 0 requires solving x = (1)/(4)tan⁻¹((1)/(2x)).
      20 mmrTotal Volume = πr²(20) + ⅔πr³ = 800 mm³⇒ πr³ + 30πr² − 1200 = 0
      Figure 6.15: Tuning peg geometry: cylinder + hemisphere volume generates π r³ + 30π r² - 1200 = 0.

      Worked Example 6.9: Stationary point of a trigonometric curve P3 only

      The curve with equation y = x² cos 4x has a stationary point P in the interval 0 < x < π/8.
      (a) Show that the x-coordinate of P satisfies x = (1)/(4)tan⁻¹((1)/(2x)).
      (b) Use the iteration xₙ₊₁ = (1)/(4)tan⁻¹((1)/(2xₙ)) with x₁ = 0.3 to find the x-coordinate of P correct to 2 decimal places.

      1. Step 1: Differentiate using the product rule:
        dy/dx = 2x cos 4x − 4x² sin 4x
      2. Step 2: At stationary point P, (dy)/(dx) = 0 ⇒ 2x cos 4x - 4x² sin 4x = 0.
      3. Step 3: Since x ≠ 0, divide by 2x cos 4x: 1 - 2x tan 4x = 0 ⇒ tan 4x = (1)/(2x).
      4. Step 4: Take the inverse tangent: 4x = tan⁻¹((1)/(2x)) ⇒ x = (1)/(4)tan⁻¹((1)/(2x)). ✔
      5. Step 5: Iterate with x₁ = 0.3:
        • x₂ = (1)/(4)tan⁻¹((1)/(2(0.3))) = (1)/(4)tan⁻¹(1.6667) = 0.2576
        • x₃ = (1)/(4)tan⁻¹((1)/(2(0.2576))) = 0.2738
        • x₄ = 0.2678, x₅ = 0.2701, x₆ = 0.2692, x₇ = 0.2696
      6. Step 6: Settle at x ≈ 0.27 (2 d.p.).
      7. Step 7: Check bounds on f(x) = 4x - tan⁻¹((1)/(2x)): f(0.265) < 0 and f(0.275) > 0. Confirmed x_P = 0.27 (2 d.p.). ✔

      ✏️ Try it

      1. A cylinder of length 20 mm has a hemispherical cap of base radius r mm. If the total volume is 800 mm³, show that π r³ + 30π r² - 1200 = 0, and verify that 3 < r < 4.
      2. A curve has equation y = x e−2x. Find the exact coordinates of its stationary point algebraically, and state why a numerical method was not required.
      3. The parametric equations of a curve are x = 2ln(t + 2), y = t³ + 2t + 3. At point P, the gradient is 0.5. Show that parameter p satisfies p = (1)/(3p² + 2) - 2.
      Answers

      1. Step 1: Volume of cylinder = π r² h = 20π r². Volume of hemisphere = (2)/(3)π r³. Step 2: Total volume = 20π r² + (2)/(3)π r³ = 800. Multiply by 3/2: 30π r² + π r³ = 1200 ⇒ π r³ + 30π r² - 1200 = 0. Step 3: Let f(r) = π r³ + 30π r² - 1200. f(3) = 27π + 270π - 1200 = 297π - 1200 ≈ -266.9 < 0; f(4) = 64π + 480π - 1200 = 544π - 1200 ≈ +509.0 > 0. Change of sign confirms 3 < r < 4.

      2. Step 1: (dy)/(dx) = e−2x + x(-2e−2x) = e−2x(1 - 2x). Step 2: Setting (dy)/(dx) = 0 ⇒ 1 - 2x = 0 (since e−2x ≠ 0), giving x = 0.5. Step 3: y = 0.5 e−1 = (1)/(2e). Step 4: Numerical methods were not needed because the derivative factorises into an elementary linear factor that solves directly.

      3. Step 1: (dx)/(dt) = (2)/(t + 2) and (dy)/(dt) = 3t² + 2. Step 2: By the chain rule, (dy)/(dx) = (dy/dt)/(dx/dt) = ((3t² + 2)(t + 2))/(2). Step 3: At P, gradient is 0.5: ((3p² + 2)(p + 2))/(2) = (1)/(2) ⇒ (3p² + 2)(p + 2) = 1. Step 4: Divide by 3p² + 2: p + 2 = (1)/(3p² + 2) ⇒ p = (1)/(3p² + 2) - 2. ✔

      Next: 6.6 Summary & quiz →

      Chapter summary: Numerical solutions of equations

      • Graphical root location: Split f(x) = 0 into g(x) = h(x). Sketch both curves on the same axes. The x-coordinates of intersection points are the roots.
      • Change-of-sign rule: If f(x) is continuous on [a, b] and f(a) and f(b) have opposite signs (f(a)f(b) < 0), at least one root lies between a and b.
      • Failure modes: The test fails or requires caution if f(x) has an asymptote (discontinuity) between a and b, or if the curve touches the axis without crossing (repeated root).
      • Iterative formula: Rearrange f(x) = 0 into x = F(x), giving recurrence relation xₙ₊₁ = F(xₙ). At the root, α = F(α).
      • Convergence criterion: The sequence converges locally if and only if |F′(α)| < 1. If |F′(α)| > 1, it diverges.
      • Cobweb vs Staircase:
        • If -1 < F′(α) < 0: iterates oscillate around α forming a convergent cobweb.
        • If 0 < F′(α) < 1: iterates move monotonically towards α forming a convergent staircase.
      • Rigorous error bounds proof: To confirm a root to d decimal places as A, evaluate f(x) at A - 0.5 × 10−d and A + 0.5 × 10−d. Opposite signs confirm that α is trapped within the rounding interval.
      Concept / MethodFormula / ConditionKey reminder
      Graphical intersectionf(x) = 0 \iff g(x) = h(x)Plot key points (intercepts, asymptotes); count intersections
      Sign-change rulef(a) × f(b) < 0 for continuous fMust explicitly state continuity and opposite signs
      Iterative formulaxₙ₊₁ = F(xₙ)Use calculator Ans key; do not round intermediate steps
      Convergence criterion|F′(α)| < 1Derivative at root must have absolute value strictly less than 1
      Cobweb pattern-1 < F′(α) < 0Iterates oscillate above and below root in an alternating spiral
      Staircase pattern0 < F′(α) < 1Iterates step monotonically from one side towards the root
      Bounds proof (d d.p.)f(A - 0.5 × 10−d) × f(A + 0.5 × 10−d) < 0Never rely only on consecutive iterates; evaluate f at bounds

      The Complete Numerical Strategy

      1. Equation f(x) = 0Graph / Sign change2. Rearrange x = F(x)Check |F′(α)| < 13. Iterate xₙ₊₁ = F(xₙ)Generate sequence4. Bounds checkTest α ± 0.5×10⁻ᵈ
      Figure 6.16: Four-stage examination workflow for solving equations numerically.

      📄 Practice questions

      All questions from Prerequisite knowledge, Explore 6.1 and 6.2, Exercises 6A, 6B and 6C, End-of-chapter review exercise 6, and Cross-topic review exercise 2 are in the printable assignment for this chapter, together with a Check yourself set and the chapter quiz.

      Mini-quiz

      1. For f(x) = x³ + 2x − 5, between which consecutive integers does the real root α lie?
      2. Which pair of graphs can be sketched to find the number of real roots of eˣ − 3x − 2 = 0?
      3. When can a sign-change check on [a, b] fail to detect a root of a continuous function f(x) = 0?
      4. Which rearrangement of x³ − 4x − 1 = 0 into the form x = F(x) is algebraically valid?
      5. Using xₙ₊₁ = √(5 − xₙ) with x₁ = 2, what is the value of x₃ correct to 4 decimal places?
      6. What condition on F′(α) guarantees that xₙ₊₁ = F(xₙ) converges for x₁ close to root α?
      7. Which diagram pattern appears when xₙ₊₁ = F(xₙ) converges with −1 < F′(α) < 0 near α?
      8. Which diagram pattern appears when xₙ₊₁ = F(xₙ) converges with 0 < F′(α) < 1 near α?
      9. The sequence xₙ₊₁ = (6 − xₙ²)/5 converges to a positive root α. What is the exact value of α?
      10. To prove that a root α of continuous f(x) = 0 is 1.47 to 2 d.p., where must a sign change be shown?
      11. Using xₙ₊₁ = 2 sin xₙ (in radians) with x₁ = 1.9, what is x₂ correct to 4 decimal places?
      12. At a stationary point of y = e²ˣ − 6x², dy/dx = 0 gives e²ˣ = 6x. Which formula isolates x from e²ˣ?

      Source: Cambridge International AS & A Level Mathematics: Pure Mathematics 2 & 3 Coursebook (Cambridge University Press, Sue Pemberton and Julianne Hughes), Chapter 6 (Numerical solutions of equations) and Cross-topic review exercise 2. Explanations, extra worked examples, Try it answers, interactive simulation, tool and quiz are Guru Jazzy's own. This page is not made or endorsed by Cambridge University Press or Cambridge Assessment International Education.