A-level Mathematics

9.3 Newton–Raphson and other recurrence relations — Practise with help

Work each question closed book first. Write down the derivative before you substitute, and keep the full value in your calculator.

Warm-up: remember

  1. Differentiate f(x)=x3−4x+1f(x) = x^3 - 4x + 1.
  2. For f(x)=x2−3xf(x) = x^2 - 3x, work out f(2)f(2) and f′(2)f'(2).
Answers
  1. f′(x)=3x2−4f'(x) = 3x^2 - 4.
  2. f(2)=4−6=−2f(2) = 4 - 6 = -2 and f′(x)=2x−3f'(x) = 2x - 3, so f′(2)=1f'(2) = 1.

If the first part was hard, revise differentiating powers in Section 7.3. If the second was hard, revise function notation in Section 2.9.

Practise with help

1. For f(x)=x3+x−1f(x) = x^3 + x - 1, write down f′(x)f'(x) and use Newton–Raphson with x0=0.7x_0 = 0.7 to find x1x_1 to four decimal places.

Hint

The step needs the derivative at the starting value, so differentiate first.

Next step

f′(x)=3x2+1f'(x) = 3x^2 + 1, so f(0.7)=0.043f(0.7) = 0.043 and f′(0.7)=2.47f'(0.7) = 2.47.

Full working f(0.7)=0.73+0.7−1=0.043,f′(0.7)=3(0.49)+1=2.47.f(0.7) = 0.7^3 + 0.7 - 1 = 0.043, \qquad f'(0.7) = 3(0.49) + 1 = 2.47. x1=0.7−0.0432.47=0.6826.x_1 = 0.7 - \frac{0.043}{2.47} = 0.6826.

2. Continue the iteration in question 1 to find x2x_2. Give the root to three decimal places.

Hint

Use x1x_1 in the same formula. Carry the full value from the last step, not the rounded one.

Next step

f(0.6825911)=0.0006312f(0.6825911) = 0.0006312 and f′(0.6825911)=2.3977918f'(0.6825911) = 2.3977918.

Full working x2=0.6825911−0.00063122.3977918=0.6823279.x_2 = 0.6825911 - \frac{0.0006312}{2.3977918} = 0.6823279.

The value has settled, so the root is 0.6820.682 to three decimal places.

3. Use Newton–Raphson with x0=2x_0 = 2 to find x1x_1 and x2x_2 for x3−2x−5=0x^3 - 2x - 5 = 0.

Hint

Differentiate f(x)=x3−2x−5f(x) = x^3 - 2x - 5 term by term.

Next step

f(2)=−1f(2) = -1 and f′(2)=10f'(2) = 10, so the first step adds 0.10.1.

Full working f′(x)=3x2−2,x1=2−−110=2.1.f'(x) = 3x^2 - 2, \qquad x_1 = 2 - \frac{-1}{10} = 2.1.

Then f(2.1)=0.061f(2.1) = 0.061 and f′(2.1)=11.23f'(2.1) = 11.23, so

x2=2.1−0.06111.23=2.0946.x_2 = 2.1 - \frac{0.061}{11.23} = 2.0946.

4. Show that xn+1=12(xn+7xn)x_{n+1} = \dfrac{1}{2}\left(x_n + \dfrac{7}{x_n}\right) is Newton–Raphson for x2−7=0x^2 - 7 = 0. Use it with x0=3x_0 = 3 to find x2x_2 to four decimal places.

Hint

Work out x−f(x)f′(x)x - \dfrac{f(x)}{f'(x)} for f(x)=x2−7f(x) = x^2 - 7, and compare it with the recurrence.

Next step

f′(x)=2xf'(x) = 2x, so x−x2−72x=x2+72xx - \dfrac{x^2 - 7}{2x} = \dfrac{x^2 + 7}{2x}.

Full working x−f(x)f′(x)=x−x2−72x=2x2−x2+72x=12(x+7x).x - \frac{f(x)}{f'(x)} = x - \frac{x^2 - 7}{2x} = \frac{2x^2 - x^2 + 7}{2x} = \frac{1}{2}\left(x + \frac{7}{x}\right).

This is the given recurrence. From x0=3x_0 = 3:

x1=12(3+73)=2.6667,x2=12(2.6667+72.6667)=2.6458.x_1 = \frac{1}{2}\left(3 + \frac{7}{3}\right) = 2.6667, \qquad x_2 = \frac{1}{2}\left(2.6667 + \frac{7}{2.6667}\right) = 2.6458.

5. Explain why Newton–Raphson fails for f(x)=x2−9f(x) = x^2 - 9 when the starting value is x0=0x_0 = 0.

Hint

Work out the derivative at the starting value.

Next step

f′(x)=2xf'(x) = 2x, so f′(0)=0f'(0) = 0.

Full working

The formula divides by f′(xn)f'(x_n). At x0=0x_0 = 0 the derivative is f′(0)=0f'(0) = 0, so the first step would divide by zero. The tangent at x0=0x_0 = 0 is horizontal and never meets the axis, so there is no crossing point to use.

6. For f(x)=x3−2x+2f(x) = x^3 - 2x + 2 with x0=0x_0 = 0, work out x1x_1, x2x_2 and x3x_3. Describe what the sequence does.

Hint

Find f(0)f(0) and f′(0)f'(0) first, then apply the formula.

Next step

f(0)=2f(0) = 2 and f′(0)=−2f'(0) = -2, so x1=0−2−2=1x_1 = 0 - \dfrac{2}{-2} = 1.

Full working x1=0−2−2=1.x_1 = 0 - \frac{2}{-2} = 1.

Then f(1)=1f(1) = 1 and f′(1)=1f'(1) = 1, so

x2=1−11=0,x3=0−2−2=1.x_2 = 1 - \frac{1}{1} = 0, \qquad x_3 = 0 - \frac{2}{-2} = 1.

The values repeat 1,0,1,0,…1, 0, 1, 0, \ldots The method does not settle, even though every step was correct.

7. For x3+x−1=0x^3 + x - 1 = 0, start both methods from x0=0.7x_0 = 0.7: Newton–Raphson and the simple iteration xn+1=1xn2+1x_{n+1} = \dfrac{1}{x_n^2 + 1}. Find x2x_2 for each. Which method is closer to the root 0.68230.6823?

Hint

The Newton–Raphson values are in question 2. For the simple iteration, substitute into 1x2+1\dfrac{1}{x^2 + 1}.

Next step

For the simple iteration, x1=10.49+1=0.6711x_1 = \dfrac{1}{0.49 + 1} = 0.6711.

Full working

Newton–Raphson reached x2=0.6823279x_2 = 0.6823279.

The simple iteration gives

x1=10.72+1=0.6711,x2=10.67112+1=0.6894.x_1 = \frac{1}{0.7^2 + 1} = 0.6711, \qquad x_2 = \frac{1}{0.6711^2 + 1} = 0.6894.

Newton–Raphson is closer after two steps. It uses the gradient as well as the height, so it closes the gap faster. The simple iteration still works, because ∣g′∣<1|g'| < 1 at the root, but it needs more steps.

Settings

Colour scheme

Applies to every book you read here.

Reading theme

Applies to every book you read here.