Thread Subject: Problem with 'lsqnonlin' function

Hello,
I am trying to do a non-linear fit using the 'lsqnonlin' Matlab function.
Some times the termination is successful and I get a message that:
Optimization terminated: relative function value changing by less than OPTIONS.TolFun.
However, there are times where I get the following termination message:
Optimization terminated: first-order optimality less than OPTIONS.TolFun,
 and no negative/zero curvature detected in trust region model.

What does it mean?
I tried looking at the Matlab help of the function but couldn't understand what this message is supposed to tell me?

Thank you

miki lo <myartsev@gmail.com> wrote in message
>
> What does it mean?
> I tried looking at the Matlab help of the function but
couldn't understand what this message is supposed to tell me?
>
> Thank you

That means MATLAB could not evaluate the gradient accurately
at the convergence precision that was asked for. It likely a
roundoff problem.

Do you provide the gradient of your cost function, or do you
let MATLAB computes it (by finite difference IIRC)?

Bruno

No, I didn't specify anything I let MATLAB compute it. Does this mean that I fit was not possible (failed)?
Also, how is this message diffrent from the message:Optimization terminated: relative function value changing by less than OPTIONS.TolFun.
because I thought the later said that the convergence was good enough so it stoped no?
Thank you

miki lo <myartsev@gmail.com> wrote in message
> Also, how is this message diffrent from the
message:Optimization terminated: relative function value
changing by less than OPTIONS.TolFun.
> because I thought the later said that the convergence was
good enough so it stoped no?

I as said before, the warning message is about gradient
evaluation that is no longer consistent with cost function
due to round-off problem. In this case, minimization
algorithm cannot go any further (it usually makes an orbit
around a local minima).

Bruno

Ok, thank you.
Just to verify, the message:
Optimization terminated: relative function value changing by less than OPTIONS.TolFun
means that the it WAS ABLE to make the fit, right?

miki lo <myartsev@gmail.com> wrote in message
> Just to verify, the message:
> Optimization terminated: relative function value changing
by less than OPTIONS.TolFun
> means that the it WAS ABLE to make the fit, right?

Yes

Bruno