Optimization toolbox-fmincon error-

Hello everybody

I'm trying to minimize a multivariable expression using fmincon under constrain. To introduce you more precisely the problem. I want to find the normal vector u of the best fitting plan which describe the most precisely a set of data from a marker placed on a rat knee. And for this I use an optimization method. But it doesn't work... My matlab version is R2012b

The expression to minimize is:

function [f] = Functiofun(S)

syms x y z
u=[x;y;z];
f=u'*S*u;

end

were S is a matrix using data of the marker and this returns:

f =

x*((4919489910730857*conj(x))/549755813888 + (1885243380871159*conj(y))/274877906944 - (1356072232583407*conj(z))/549755813888) - z*((1356072232583407*conj(x))/549755813888 + (4128876282184185*conj(y))/2199023255552 - (3333614421791133*conj(z))/137438953472) + y*((1885243380871159*conj(x))/274877906944 + (924159128957421*conj(y))/68719476736 - (4128876282184185*conj(z))/2199023255552)

Hence I want to minimize this function under the following constrain:

function [g] = Functionfun_g(C)

syms x y z
u=[x;y;z];
g=u'*C*u-1;

end

which is simply normalisation of my output vector

g =

x*conj(x) + y*conj(y) + z*conj(z) - 1

so I wrote the following code sentence using fmincon:

x0=[0;0;0]
options = optimset('Algorithm','sqp');
[x,fval] = fmincon(f,x0,[],[],[],[],[],[],...
g,options);

And it returns me this error:

      Error using optimfcnchk (line 288)
      If FUN is a MATLAB object, it must have an feval method.

      Error in fmincon (line 423)
         funfcn =
         optimfcnchk(FUN,'fmincon',length(varargin),funValCheck,flags.grad,flags.hess,false,Algorithm);

.... I need help please :)

Answer 1

[Igor]

You can't pass symbolic variables as functions to 'fmincon'. Use 'matlabFunction' to convert your symbolic functions to function handles: http://www.mathworks.com/help/symbolic/matlabfunction.html

Good luck

Answer 2

[Victor Moranville]

Hello

I found the mistake and as you said it was a function problem declaration I wrote it as following:

[S,C]=Matrix_C_S(D)
options = optimset('Algorithm','sqp','MaxIter',2000);
x0=[1;1;1;1;1]
[x, k] = fmincon(@(x)([x(1),x(2),x(3),x(4),x(5)]*(S*[x(1);x(2);x(3);x(4);x(5)])),x0,[],[],[],[],[],[],...
@confuneq,options);options = optimset('Algorithm','sqp','MaxIter',2000);
x

;)

Danke Shoën