Additional output with fsolve

14 vues (au cours des 30 derniers jours)
Theo
Theo le 14 Mai 2011
I'm asking what seems to be an identical question to this one. Specifically, I simply want to output additional variables after fsolve completes its task. Suppose that I am using fsolve with this function:
function [F G] = myfunc(x)
F = x^2;
G = 2;
I will be using fsolve with
xsol = fsolve(@(x)myfunc(x), guess);
This will give me the solution. However, if I want the value of G, I need to call myfunc again
[F, G] = myfunc(xsol)
I'd like to NOT evaluate myfunc again. How do I get fsolve to also output additional variables? I'd like to do so without global variables or outputting to files...

Réponse acceptée

Andrew Newell
Andrew Newell le 14 Mai 2011
Is the G you're really calculating costly to evaluate? If so, I'd suggest a different approach. Define your function as
function [F G] = myfunc(x)
F = x^2;
if nargout > 1
G = 2;
end
so it doesn't evaluate G during the solving (which would involve, generally, several calls to myfunc ). Note that you can use a more compact notation:
xsol = fsolve(@myfunc,guess);
Then go ahead and calculate G once:
[F G] = myfunc(x)
You'll probably save a lot more computation with this approach.
  2 commentaires
Theo
Theo le 14 Mai 2011
I'm sorry I was not more specific (I tried to simplify the problem as much as I could).
> Is the G you're really calculating costly to evaluate?
Sort of. But it is also needed to determined F (and hence for fsolve). Think of G is a function and F as its norm (or some other property). Thus, fsolve will solve for G as it solves F = 0.
Andrew Newell
Andrew Newell le 14 Mai 2011
I do appreciate your obvious effort to distill the essence of this problem.

Connectez-vous pour commenter.

Plus de réponses (1)

Andrew Newell
Andrew Newell le 14 Mai 2011
Based on your comments, here is another approach. Define a function
function G = myGfunc(x)
persistent G
if nargin > 0
% calculate G
end
This saves the value of G you calculate. Then define
function [F G] = myfunc(x)
G = myGfunc(x);
F = GtoF(G); % Your method of calculating F using G
After your run, you can enter
G = myGfunc;
and it will simply return the saved value of G.
Caveat: it would be wise to enter
clear functions
before each run to make sure that G isn't saved from a previous run.

Tags

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by