0
$\begingroup$

I'm working on a nonlinear system of 2 equations and try to solve it with the Newton-Raphson method. I have a function $f(x,y)$. In order to use the method, I have to find a starting vector $(x,y)$. But the method will only converge for specific starting vectors.

Is there a formula that gives the convergence of the Newton-Raphson method on a system of equations ?

Thank you all.

$\endgroup$
0
$\begingroup$

As long as you pick even a somewhat reasonable guess, the method should find the closest root. If the equations were linear, and you never encounter a situation where you divide by zero, you won't even have to use a reasonable guess.

$$ {\displaystyle x_{n+1}=x_{n}-m{\frac {f(x_{n})}{f'(x_{n})}}.} $$

$\endgroup$
1
  • $\begingroup$ The problem is that I use the N-R method in a program and I don't know some parameters of my function $f(x,y)$. That means I cannot make a "reasonable guess". What I do know is that $f(x,y)$ is nonlinear. I would need a formula that determines wether the method converges in terms of the function $f(x,y)$. $\endgroup$ – Luc Dec 5 '18 at 16:51
0
$\begingroup$

No, in general there is no easy way to determine whether or not Newton-Raphson will converge for a given starting value, other than to try it. The region of convergence can be very complicated, and may have a fractal boundary.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.