Re: Non linear equations: howto?

[Ben Abbott]

On Oct 20, 2008, at 12:47 AM, Fabrice DELENTE wrote:

> > > My problem is that I have the curve of f, and I'd like to find its  
> > > equation;
> > > I guessed that f is of the form (a*x+b)*exp(c*x+d)+e, and I'm  
> > > trying to find
> > > the right values for a,b,c,d and e.
>
> > This is impossible to solve. E.g. d can be anything, so I suggest  
> > making
> > it 0.
>
> Well I tried on geogebra to find a solution visually, and it seems  
> possible;
> why do you say it is impossible?

You problem requires a non-linear solution of six eqns for 5 unknowns.

There can be many locally minimal solutions, and (generally) no way to  
determine how many local minima exist or where they are.

Ben