9709 P3 §3.6 Numerical Solution of Equations

Sign Change and Fixed-Point Iteration
A-Level Mathematics · §3.6 Numerical Solution of Equations
💡 Lead In

Looks Simple — Can You Solve It?

Take $x^{3}-x-2=0$ and try to solve it exactly. Factorising turns up no pair that works, and there is no cubic formula you would reach for in an exam. So a fair question: does this equation even have a solution — and if it does, how can we pin it down?

⚠️
Watch out! A sign change only traps a root inside an interval — it does not hand you its value (it is not simply the midpoint). And not every rearrangement will close in on the root.
→
Today's goal. Trap a root with a sign change, rearrange $f(x)=0$ into $x=F(x)$, then iterate $x_{n+1}=F(x_n)$ until the values settle — and judge whether a given rearrangement converges.

Three lines of reasoning below use these ideas. Predict each one first, then click Correct step or Wrong step.

Step 1 — $f(1)=-2$ and $f(2)=4$ have opposite signs, so the curve must cross the axis in $(1,2)$.
Step 2 — because $f(1)\lt 0\lt f(2)$, the root is exactly $x=1.5$, the midpoint of the interval.
Step 3 — from $x^{3}=x+2$ we get $x=\sqrt[3]{x+2}$, so $x_2=\sqrt[3]{1.5+2}=1.51829$.

Those two moves answer the opening question. The sign change traps the root in $(1,2)$; feeding each answer back through the rearrangement gives $x_1=1.5,\ x_2=1.51829,\ x_3=1.52094,\ x_4=1.52132,\ldots$, which homes in on the root $1.521$ — accurate enough to use.

🧠 Activating Prior Knowledge

Try each in your head before clicking an answer — this is a test, not a reveal. See how many you get first try. The clock starts on your first answer.

R1. A root of $f(x)=0$ is a value of $x$ where:

the curve y=f(x) crosses the x-axis
the curve y=f(x) has a maximum
the curve y=f(x) crosses the y-axis
the gradient of the curve is zero
Choose an option above.
A root is an $x$-axis crossing: it solves $f(x)=0$.

R2. The function $f$ is continuous with $f(1)=-2$ and $f(2)=5$. What follows?

nothing at all
a root of f(x)=0 lies in (1,2)
the root is exactly 1.5
f has two roots in (1,2)
Choose an option above.
A change of sign across a continuous function guarantees at least one root in the interval.

R3. Make $x$ the subject of $x^{3}=x+2$:

$x=x+2$
$x=\sqrt[3]{x+2}$
$x=x^{3}-2$
$x=(x+2)^{3}$
Choose an option above.
Take the cube root of both sides: $x=\sqrt[3]{x+2}$.
📖 Key Concepts
Explore 1 — the sign-change criterion.

If $f$ is continuous on $[a,b]$ and $f(a)$ and $f(b)$ have opposite signs, then $f(x)=0$ has at least one root in $(a,b)$:

$f(a)\lt 0\lt f(b)\ \Rightarrow\ \exists\ c\in(a,b)\text{ with }f(c)=0$

Evaluate at consecutive integers and look for the change of sign. For $f(x)=x^{3}-x-2$:

x0123
f(x)-2-2422

The sign changes between $1$ and $2$, so a root lies in the unit interval $(1,2)$. A sign change only traps the root — it does not yet tell us its value.

Continuity is essential: a function that jumps over the axis can change sign without ever crossing it.

Explore 2 — rearrange $f(x)=0$ into $x=F(x)$, then iterate.

From $x^{3}-x-2=0$ we get $x^{3}=x+2$, and taking cube roots gives:

$x=\sqrt[3]{x+2}$

Now feed each answer back into the same formula, $x_{n+1}=F(x_n)=\sqrt[3]{x_n+2}$, starting from $x_1=1.5$:

n12345
xₙ1.51.518291.520941.521321.52137
xₙ₊₁1.518291.520941.521321.521371.52138

The iterates settle on $x=1.521\ (3\ \text{d.p.})$.

A value that satisfies $x=F(x)$ is a fixed point of $F$ — and it is exactly a root of the original equation, because the two equations are equivalent.

Explore 3 — convergence depends on the rearrangement.

A different rearrangement of $x^{3}-x-2=0$ is $x=x^{3}-2$, i.e. $x_{n+1}=x_n^{3}-2$. From $x_1=1.5$:

$x_1=1.5,\ x_2=1.375,\ x_3=0.600,\ x_4=-1.784,\ldots$

The values move apart, so this iteration diverges and never reaches the root. In general the iteration converges when the iterates settle down; near the root this happens when $|F'(x)|\lt 1$.

Group Iteration Explorer: converge or diverge

Pairs pick a starting value x₁, then press Iterate to feed each answer back through x_{n+1} = F(x_n). Use the buttons to switch between two rearrangements of x³ − x − 2 = 0: the live plot of y = F(x) with y = x, and the table, show the iterates stepping toward the fixed point for one rearrangement and away from it for the other. Take turns: one partner sets the start and presses Iterate, the other predicts the next value first.

Individual Spot the Error: rearrangements and iteration

Each card shows a sign-change claim, a rearrangement or an iteration result. Decide whether it is correct, then explain the slip out loud to a partner.

✏️ Worked Examples
Exercise 6B Q1 Worked Example

(i) Show that $x^{3}-x-2=0$ has a root between 1 and 2. (ii) The equation is rearranged into the form $x=F(x)$ with $F(x)=\sqrt[3]{x+2}$. Use the resulting iterative formula to find the root correct to 3 decimal places.

Hint

Trap the root with a sign change first. Then feed each answer back into $F(x)=\sqrt[3]{x+2}$ — never back into the original start.

Step 1 — a root in (1, 2)

$f(1)=1-1-2=-2\lt 0$ and $f(2)=8-2-2=4\gt 0$, so by the sign-change criterion a root lies in $(1,2)$.

Step 2 — the iterative formula

From $x^{3}=x+2$, $x=\sqrt[3]{x+2}$, so $x_{n+1}=\sqrt[3]{x_n+2}$.

Step 3 — iterate from $x_1=1.5$

$x_1=1.5$
$x_2=1.51829$
$x_3=1.52094$
$x_4=1.52132$
$x_5=1.52137$

which has settled at

$x=1.521\ (3\ \text{d.p.})$

Exercise 6B Q4 Worked Example

(i) Show that $\mathrm{e}^{x}+x-6=0$ has a root in the interval $[1,2]$. (ii) Show that the equation may be written $x=\ln(6-x)$. (iii) Use an iterative formula based on this equation to find the root correct to 3 decimal places.

Hint

Rearrange to isolate $\mathrm{e}^{x}$, then take natural logs of both sides to free the $x$.

Step 1 — a root in [1, 2]

$f(1)=\mathrm{e}+1-6\approx -2.282\lt 0$ and $f(2)=\mathrm{e}^{2}+2-6\approx 3.389\gt 0$, so a root lies in $(1,2)$.

Step 2 — the log rearrangement

$\mathrm{e}^{x}=6-x\ \Rightarrow\ x=\ln(6-x)$, giving $x_{n+1}=\ln(6-x_n)$.

Step 3 — iterate from $x_1=1.5$

$x_1=1.5$
$x_2=1.50408$
$x_3=1.50317$
$x_4=1.50337$
$x_5=1.50333$

so

$x=1.503\ (3\ \text{d.p.})$

📝 Concept Quiz

Answer, then check. Options are shuffled every load. The clock starts on your first answer.

Q1. A rearrangement of $x^{3}-8x-13=0$ into the form $x=F(x)$ is:

$x=\sqrt[3]{8x+13}$
$x=\sqrt[3]{8x-13}$
$x=\sqrt[3]{13-8x}$
$x=(8x+13)^{3}$
Choose an option above.
$x^{3}=8x+13$, so $x=\sqrt[3]{8x+13}$ — the $8x$ keeps its plus sign.

Q2. The first iterate of $x_{n+1}=\sqrt[3]{x_n+2}$ from $x_1=1.5$ is:

$1.500$
$1.518$
$1.521$
$3.500$
Choose an option above.
$x_2=\sqrt[3]{1.5+2}=\sqrt[3]{3.5}=1.518$ (3 d.p.).

Q3. Iterating $x_{n+1}=x_n^{3}-2$ from $x_1=1.5$:

converges to 1.521
converges to 1.5
converges to -1.784
diverges — the values move apart
Choose an option above.
The values $x_1=1.5,\ x_2=1.375,\ x_3=0.600,\ x_4=-1.784,\ldots$ move apart, so this rearrangement diverges.

Q4. The equation $x=\cos x$ ($x$ in radians) is solved by iterating $x_{n+1}=\cos x_n$. Starting from $x_1=0.7$, the root correct to 3 d.p. is:

$0.739$
$0.700$
$0.764$
$0.540$
Choose an option above.
The iterates $x_1=0.70000,\ x_2=0.76484,\ x_3=0.72149,\ldots$ settle at $0.73909$, so the root is $0.739$ (3 d.p.).
📚 More Practice

Work each on paper, then reveal the solution to check your working.

Exercise 6B Q3 Standard

(i) Show that $\mathrm{e}^{-x}-x+2=0$ has a root in the interval $[2,3]$. (ii) The equation is rearranged as $x=\mathrm{e}^{-x}+2$. Use the resulting iterative formula to find the root correct to 3 decimal places.

Reveal Solution

$f(2)=\mathrm{e}^{-2}\approx 0.135\gt 0$ and $f(3)=\mathrm{e}^{-3}-1\approx -0.950\lt 0$, so a root lies in $(2,3)$.

Iterating $x_{n+1}=\mathrm{e}^{-x_n}+2$ from $x_1=2.5$:

$x_1=2.5$
$x_2=2.08208$
$x_3=2.12467$
$x_4=2.11947$
$x_5=2.12009$
$x_6=2.12002$

$x=2.120\ (3\ \text{d.p.})$

Exercise 6B Q14 Medium

The equation $x^{3}-8x-13=0$ has one real root. (i) Find the two consecutive integers between which this root lies. (ii) Use the iterative formula $x_{n+1}=\sqrt[3]{8x_n+13}$ to find this root correct to 2 decimal places, giving each iteration to 4 decimal places.

Reveal Solution

$f(3)=27-24-13=-10\lt 0$ and $f(4)=64-32-13=19\gt 0$, so the root lies between $3$ and $4$.

Iterating from $x_1=3.5$:

$x_1=3.5$
$x_2=3.4482$
$x_3=3.4366$
$x_4=3.4339$
$x_5=3.4333$
$x_6=3.4332$

$x=3.43\ (2\ \text{d.p.})$

Exercise 6B Q12 Hard

The curve $y=x\mathrm{e}^{2x}$ has a minimum point $M$. (i) Find the exact coordinates of $M$. (ii) Show that the curve meets the line $y=20$ at the point whose $x$-coordinate is a root of $x=\dfrac{1}{2}\ln\!\left(\dfrac{20}{x}\right)$. (iii) Use the iterative formula $x_{n+1}=\dfrac{1}{2}\ln\!\left(\dfrac{20}{x_n}\right)$ from $x_1=1.3$ to find the root correct to 2 decimal places.

Reveal Solution

$\dfrac{dy}{dx}=\mathrm{e}^{2x}(1+2x)=0$ at $x=-\dfrac{1}{2}$, giving $M\left(-\dfrac{1}{2},\ -\dfrac{1}{2\mathrm{e}}\right)$.

Setting $x\mathrm{e}^{2x}=20$ gives $\mathrm{e}^{2x}=\dfrac{20}{x}$, and taking logs: $x=\dfrac{1}{2}\ln\!\left(\dfrac{20}{x}\right)$.

Iterating from $x_1=1.3$:

$x_1=1.3$
$x_2=1.3667$
$x_3=1.3417$
$x_4=1.3509$
$x_5=1.3475$
$x_6=1.3487$

$x=1.35\ (2\ \text{d.p.})$

🎫

Exit Ticket

Check the learning objectives you feel confident about: