Re: Return additional output in optimization

On 4/7/2016 10:11 AM, Nana wrote:
> Alan Weiss <aweiss@mathworks.com> wrote in message
> <ne5m6f$k02$2@newscl01ah.mathworks.com>...
>> On 4/7/2016 7:39 AM, Nana wrote:
>> > I want to optimize p from testfun(). However I want to return
>> additional
>> > output variable (b) from the optimization.
>> >
>> > Function [p, b] = testfun()
>> >
>> > x = fmincon(@testfun, x0,options)
>> >
>> > where b is the additional variable passed to fsolve, I am looking to
>> > achieve something like this,
>> > [x,b] = fmincon(x0,options)
>> >
>> > How can I do that?
>> > Thx
>> > Nana
>>
>> The easiest way might be to use a nested function along these lines:
>> http://www.mathworks.com/help/optim/ug/passing-extra-parameters.html#brhkghv-10
>>
>>
>> Alan Weiss
>> MATLAB mathematical toolbox documentation
>
> Hi Alan,
>
> Can you give an explicit example to do it using nested function? Because
> on the link you gave me only return x and fval as outputs.

Modify the listed example to return another output:

function [x,fval,newnew] = runnested2(a,b,c,x0)
newnew = 0; % allocation
[x,fval] = fminunc(@nestedfun,x0);
% Nested function that computes the objective function
     function y = nestedfun(x)
         y = (a - b*x(1)^2 + x(1)^4/3)*x(1)^2 + x(1)*x(2) +...
             (-c + c*x(2)^2)*x(2)^2;
         newnew = y^2 + y;
     end
end

Try it out:
a = 4; b = 2.1; c = 4;% Assign parameter values
x0 = [0.5,0.5];
[x,fval,newoutput] = runnested2(a,b,c,x0)

Alan Weiss
MATLAB mathematical toolbox documentation