Call Us 07766496223
We can find an expression for the kth error  
\[\epsilon_k\]
  of the Newton Raphson method, which estimates a root  
\[\alpha\]
  of the equation  
\[f(x)=0\]
  using the iteration formula  
\[x_{k+1}=x_k- \frac{f(x_k)}{f'(x_k)}\]
.
We can write this as  
\[x_{k+1}-x_k=- \frac{f(x_k)}{f'(x_k)}\]
  (1)
Write  
\[x_{k+1}-\alpha=x_k-\alpha+(x_{k+1}-x_k)\]

\[\epsilon_{k+1}=\epsilon_k+(x_{k+1}-x_k)\]

Into this equation substitute (1) to give  
\[\epsilon_{k+1}=\epsilon_k- \frac{f(x_k)}{f'(x_k)}\]

Now expand numerator and denominator of the second term in Taylor series about  
\[\alpha\]
.
\[\begin{equation} \begin{aligned} \epsilon_{k+1} &= \epsilon_k- \frac{f(\alpha)+f'(\alpha)(x_k-\alpha)+\frac{1}{2}f''(\alpha)(x_k-\alpha)^2+ ...}{f'(\alpha)+f''(\alpha)(x_k-\alpha)+\frac{1}{2}f''''(\alpha)(x_k-\alpha)^2+ ...} \\ & \simeq \epsilon_k- \frac{f'(\alpha)\epsilon_k+\frac{1}{2}f''(\alpha) \epsilon_k^2}{f'(\alpha)+f''(\alpha)\epsilon_k+\frac{1}{2}f''''(\alpha)\epsilon_k^2} \\ &= \epsilon_k- \epsilon_k \frac{f'(\alpha)+\frac{1}{2}f''(\alpha) \epsilon_k}{f'(\alpha)+f''(\alpha)\epsilon_k+\frac{1}{2}f''''(\alpha)\epsilon_k^2} \\ &\simeq \epsilon_k- \epsilon_k \frac{f'(\alpha)+\frac{1}{2}f''(\alpha) \epsilon_k}{f'(\alpha)+f''(\alpha)\epsilon_k+\frac{1}{2}f''''(\alpha)\epsilon_k^2} \\ &= \epsilon_k - \frac{f''(\alpha)}{2f'(\alpha)} \epsilon_k^2 \end{aligned} \end{equation}\]

Convergence is quadratic.