Nonlinear Equation Solve

[Dmitri A. Sergatskov]

---------- Forwarded message ----------
From: Dmitri A. Sergatskov <address@hidden>
Date: Tue, Feb 14, 2012 at 3:17 PM
Subject: Re: Nonlinear Equation Solve
To: "Thomas D. Dean" <address@hidden>


On Tue, Feb 14, 2012 at 2:22 PM, Thomas D. Dean <address@hidden> wrote:

>   fsolve (@eqnsetf,y',optimset("TolX",1e-20,"TolFun",1e-20));
>> info
> info = 0
>
> If I reduce the tol's, fsolve does not reach a solution.
>
>> [x, fval, info] =
>    fminunc (@eqnsetf,y',
>             optimset("TolX",1e18,"TolFun",1e18,"ComplexEqn","on"));
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

This is kind of too radical (or did you actually used 1e-18?).
Try 1e-4 or may be 1e-8

Tolerances that is much smaller than eps look suspicious to me.

> Tom Dean

Dmitri.
--