Re: Solver Question

[Jim Martin <jim martin utah edu>]

Hello Mario:

Yes, the process I describe was certainly overkill for a simple linear 
function. I tried that only as a simple example. The context I use it in 
in the manuscript I described is considerably more challenging. I use it 
to determine coefficients in a 10th order Fourier Series Approximation. 
I don't believe that linest would help me much with that.

Since posting this to the gnumeric list, I also posted it to the 
OpenOffice Calc forum. There I got a response from someone who directed 
me to a nonlinear solver extension. I have loaded that in to Calc and it 
works fine.

Thanks,

Jim

gnumeric-list-request gnome org wrote:
> Date: Mon, 23 Mar 2009 17:21:02 -0500
> From: Mario Rappi <mrappi gmail com>
> Subject: Re: Solver Question
> To: gnumeric-list gnome org
>
> Jim,
>
> I think I was able to reproduce the problem: Solver wants constraints. Even
> tried adding a dummy constrained variable. Obviously the solver engine
> doesn't like that.
> Using Solver for this seems an overkill anyway, why don't you try regression
> functions: =SLOPE(range of y's, range of x's) and =INTERCEPT(range of y's,
> range of x's)
>
> Good luck,
>
> Mario
>
>   

--
James C. Martin PhD
Associate Professor
The Neuromuscular Function Lab
Department of Exercise and Sport Science
College of Health
The University of Utah
250 S. 1850 E. Room 241
Salt Lake City, Utah, 84112-0920
Voice 801-587-7704
Fax 801-585-3992
http://www.health.utah.edu/ess/people/martin.html
"The best way to have a good idea is to have a lot of ideas." Linus Pauling