Prev: linear model from nonlinear
Next: Parsing a filename
From: Kevin on 17 Nov 2009 09:46
Hello,

I have a question to which the answer is most likely simple. I'm having trouble using anonymous functions with fsolve. The gist of my problem is this:

I have a somewhat large function that I would like to break into smaller functions, all of which are depenedent on one variable (say, x in the example below). Ultimately, I would like to use fsolve to find the root of the function with respect to x. However, I get the following error message:

"?? Undefined function or method 'plus' for input arguments of type 'function_handle'.
Error in ==> @(x)a+b-x
Error in ==> fsolve at 254
fuser = feval(funfcn{3},x,varargin{:});
caused by:
Failure in initial user-supplied objective function evaluation. FSOLVE cannot continue."

My actual function is much more complex, but the series of functions below layout the same basic structure. I believe that the error is syntax in nature, but I can not seem to find the proper construction. Any help would be appreciated.

a=@(x)x^2+2
b=@(a)a+5
c=@(x)a+b-x
fsolve(c,1)
From: Steven Lord on 17 Nov 2009 10:21

"Kevin " <remove_underscores_kevin_west(a)southalabama.edu> wrote in message
news:hducvb$sib$1(a)fred.mathworks.com...
> Hello,
>
> I have a question to which the answer is most likely simple. I'm having
> trouble using anonymous functions with fsolve. The gist of my problem is
> this:
>
> I have a somewhat large function that I would like to break into smaller
> functions, all of which are depenedent on one variable (say, x in the
> example below). Ultimately, I would like to use fsolve to find the root
> of the function with respect to x. However, I get the following error
> message:
>
> "?? Undefined function or method 'plus' for input arguments of type
> 'function_handle'.

That's correct. You can't add together function handles; what you want to
do is add together _the values returned by those function handles_.

> Error in ==> @(x)a+b-x
> Error in ==> fsolve at 254
> fuser = feval(funfcn{3},x,varargin{:});
> caused by:
> Failure in initial user-supplied objective function evaluation. FSOLVE
> cannot continue."
>
> My actual function is much more complex, but the series of functions below
> layout the same basic structure. I believe that the error is syntax in
> nature, but I can not seem to find the proper construction. Any help
> would be appreciated.
>
> a=@(x)x^2+2
> b=@(a)a+5
> c=@(x)a+b-x

a and b are function handles, which as I've said above you can't add
together. What you can do is:

a = @(x) x.^2+2;
b = @(x) a(x)+5;
c = @(x) a(x)+b(x)-x;
fsolve(c, 1)

But note that as written, this doesn't have a solution over the reals; you
can confirm this by looking at the graph and zooming in on the minimum.

ezplot(c)

or by expanding it out:

a = x^2+2
b = (x^2+2)+5 = x^2 + 7
c = (x^2+2)+(x^2+7)-x = 2*x^2-x+9

and using ROOTS:

roots([2 -1 9])

--
Steve Lord
slord(a)mathworks.com
comp.soft-sys.matlab (CSSM) FAQ: http://matlabwiki.mathworks.com/MATLAB_FAQ


 | 
Pages: 1
Prev: linear model from nonlinear
Next: Parsing a filename