The convergence of the fixed-point iteration for solving a cubic equation

886 Views Asked by At

I have a third-grade polynomial of the form $Ax^3+Bx+C$ and I want to find its roots. I cannot use Gauss to guess the first root and it is not trivial, so I try this:

$0=Ax^3+Bx+C$

and for a given root $X_0$:

$X_0=(-AX_0^3-C)/B$

I then guess that $X_0 = 3$ (just an example), then find a new value of $X$ until it converges to something near $3.54$. Sometimes when I'm not lucky it diverges.

I assume there must be a criterion to decide whether or not $X_0$ is a good approximation in the sense that it will converge to the root.