Newton's Method & Related Methods

Overview

Newton's Method is a powerful technique for finding roots of equations. Its iterative formula is:

\( x_{n+1} = x_n - \frac{f(x_n)}{f'(x_n)} \)

Other related methods include the secant, bisection, and false position methods. This page focuses on Newton's Method and provides an interactive example.

Additional Notes

Secant Method

Use if \(f'(x)\) is unavailable.
Iterative formula: \( x_{n+1} = x_n - f(x_n) \cdot \frac{x_n - x_{n-1}}{f(x_n) - f(x_{n-1})} \).
Instructions: Start with two initial guesses and repeat until convergence.

Bisection Method

Use when \(f(a)\) and \(f(b)\) have opposite signs.
Process:

  1. Calculate midpoint: \( c = \frac{a+b}{2} \).
  2. If \( f(c) \) has the same sign as \( f(a) \), set \( a = c \); otherwise, set \( b = c \).
  3. Repeat until the interval \( [a, b] \) is sufficiently small.

Convergence is linear.

False Position (Regula Falsi) Method

Similar to bisection but uses a secant line.
Iterative formula: \( c = b - f(b) \cdot \frac{b-a}{f(b)-f(a)} \).
Instructions: Update the interval by replacing the endpoint that shares the sign with \( f(c) \).

Order of Convergence

To estimate the order \( p \) of convergence, if you have errors \( e_{n-1}, e_n, e_{n+1} \) from successive iterations, use:

\( p \approx \frac{\ln\left(|e_{n+1}/e_n|\right)}{\ln\left(|e_n/e_{n-1}|\right)} \)

This gives you an idea of how fast the error decreases (e.g., \( p \approx 2 \) for quadratic convergence and \( p \approx 1 \) for linear).