Source
Adapted from my own handwritten and hand-typed notes from Year 1 to Year 3 of my EEE degree
The Newton-Raphson method finds a root of by following tangent lines. It is simple and converges very fast when the starting guess is close to the root.
Formula
The tangent at has slope . Setting it to cross the axis gives the next guess:
This is fixed-point iteration with .
Cautions:
- if the step fails, so pick another
- it only converges reliably if is close enough to , so choose inside a bracket with
Method
Steps:
- Find with
- Pick close to the root and iterate the formula
- Stop when , with
Worked example
Find the root of in to 4 decimal places.
Solution:
- , so and we stop when
| 0 | 1 | 1 |
| 1 | 0.75 | 0.1719 |
| 2 | 0.686047 | 0.00894 |
| 3 | 0.682340 | 0.0000282 |
So to 4 decimal places after only 3 steps. Compare with fixed-point iteration, which needs many more.
Second example
Positive solution of : use , . Since and , the root is in , where . Starting at gives .
Why it is so fast is explained in Order of Convergence.