Is the command ''fzero'' based on Newton Raphson Method?

I would like to know whether the function fzero is based on Newton-Raphson method, if so then I don't have to write extra code for the NR algorithm.

Answer 1

[Mischa Kim]

Karthik, the short answer is no. fzero is based on a combination of the bisection, secant, and inverse quadratic interpolation methods. See the documentation (at bottom of page) for more detail.

[Matt J]

Hmmm, makes you wonder whether FZERO requires differentiability of the function. The bisection part does not, but the secant and quadratic interpolation parts might... Don't see anything in the documentation addressing that.

[John D'Errico]

I'd need to look at the code, but in general these types of codes are set up so that when the function is well behaved (sufficiently smooth), they will use a higher order method that would presume differentiability. If that fails however, they back off and revert to a lower order scheme (like bisection) that will be more robust to problems.