Failure in initial objective function evaluation. FMINCON cannot continue. error in fmincon line(536)

I am trying to optimize my Gibbs energy function for a mixture of gases with fmincon and I keep getting 3 errors. my function looks like this.

function G=myGfunc(nj)
Enj=sum(nj);
G=sum(nj.*(Gjo/R/T+ log(nj/Enj*(p/po))));
end

and I call the fmincon as follows:

options=optimset('Algorithm', 'interior-point');
x=fmincon(@myGfunc,x0,[],[],Aeq,beq,LB,[],[],options);

the errors I get are:

Undefined function or variable 'Gjo'.

Error in myGfunc (line 3)
G=sum(nj.*(Gjo/R/T+ log(nj/Enj*(p/po))));

Error in fmincon (line 536)
      initVals.f = feval(funfcn{3},X,varargin{:});

Error in myMeth (line 70)
  x=fmincon(@myGfunc,x0,[],[],Aeq,beq,LB,[],[],options);

Caused by:
    Failure in initial objective function evaluation. FMINCON cannot continue.

It's weird since my Gjo is well defined and it gave solid values, before not defining the parameters needed for fmincon.

any help would by greatly appreciated

Answer 1

[Walter Roberson]

Your Gjo is initialized to 8 x 1, but in your function you use Gjo' which makes that into 1 x 8. You did not happen to notice that because you are using R2016b or later, which has "implicit expansion", and treated the addition of an 8x1 with a 1x8 by doing the equivalent of bsxfun(@plus, 8x1, 1x8) giving an 8 x 8 result.

Files attached. The one to run is abs_entrop

[Sarv]

it finally worked. it was a great great help. thank you very much.

Answer 2

[Torsten]

I suspect there is an argument for the log-function which is negative or undefined.

Best wishes

Torsten.

[Sarv]

That was also true I modified it but still it won't work!

Answer 3

[Alan Weiss]

Almost certainly, you are not passing the value of Gjo that is in your workspace. See Passing Extra Parameters.

Alan Weiss

MATLAB mathematical toolbox documentation

[Sarv]

that's true. I defined it as global , but still I am getting a series of new errors!

here's my function now:

    function G=myGfunc(nj)
    global T R P Po Gjo 
    Enj = sum(nj);
    h = P/Po*nj/Enj;
    w = log(h);
    w(h==0) = 0;
    G = sum(nj.*(Gjo'+ R*T* w));
    end

and I my scrip:

Aeq = [0   1    2    0   1  0  2  0    % oxygen balance
       4   0    0    2   2  0  0  0    % hydrogen balance
       1   1    1    0   0  0  0  1    % carbon balance
       0   0    0    0   0  2  0  0];  % nitrogen balance

     %the mole values in the feed

     beq = [8  % moles of oxygen atoms coming in
            32   % moles of hydrogen atoms coming in
            4     % moles of carbon atoms coming in
            2.2];  % moles of nitrogen atoms coming in

    %Constraints on bounds

    LB=[0 0 0 0];
    %solve the minimization problem
    x0 = [1e-1; 1e-1; 1e-1; 1e-1; 1e-1; 1e-1; 1e-1; 1e-1]; % initial guess

    options=optimset('Algorithm', 'interior-point');
    x=fmincon(@myGfunc,x0,[],[],Aeq,beq,LB,[],[],options);

and finally the errors:

    Warning: Length of lower bounds is < length(x); filling in missing lower bounds with -Inf. 
    > In checkbounds (line 33)
      In fmincon (line 308)
      In myMeth_081017 (line 66) 
    Error using fmincon (line 609)
    Supplied objective function must return a scalar value.

    Error in myMeth_081017 (line 66)
      x=fmincon(@myGfunc,x0,[],[],Aeq,beq,LB,[],[],options);

I truly appreciate ur help.

[Walter Roberson]

You did not initialize your global variables. All of those global variables are going to be [] because they have not been initialized. Your routine is going to be returning [] rather than a scalar.

[Sarv]

oh yes I did. I just did not post them here. here is how I initialized them:

global T R P Po Gjo 
T=1000; %K
R=8.314e-3; 
P=6; %bar, 
Po=1; %bar,
species= {'CH4' 'CO' 'CO2' 'H2' 'H2O' 'N2' 'O_2' 'C'};

Hf298= [-74; -110.53; -393.51; 0.0; -241.826; 0.0; 0.0; 0.0];

WB = [-0.703029  108.4773  -42.52157  5.862788    0.678565  -76.84376   158.7163   -74.87310 %CH4
      25.56759  6.096130   4.054656   -2.671301   0.131021  -118.0089   227.3665   -110.5271  % CO
      24.99735  55.18696   -33.69137   7.948387   -0.136638 -403.6075   228.2431   -393.5224   % CO2
      33.066178 -11.363417  11.432816  -2.772874  -0.158558  -9.980797  172.707974  0.0          %H2
      30.09200  6.832514    6.793435   -2.534480  0.082139   -250.8810  223.3967   -241.8264    % H2O
      19.50583  19.88705   -8.598535  1.369784    0.527601   -4.935202  212.3900   0.0          %N2
      31.32234  -20.23531   57.86644   -36.50624  -0.007374  -8.903471  246.7945     0.0         %O2
      21.17510  -0.812428   0.448537   -0.043256  -0.013103   710.3470  183.8734    716.6690];  

t=T/1000;
T_H = [t; t^.2/2; t.^3/3; t.^4/4; -1./t; 1; 0; -1];
T_S = [log(t); t; t.^2/2; t.^3/3; -1./(2*t.^2); 0; 1; 0];

H=WB*T_H; %(H-H_298.15)
S=WB*T_S/1000; %absolute entropy kj/mol/k

Gjo= Hf298+H-T*S;

Answer 4

[Alan Weiss]

I suggest that you use the debugger and see what the size of the returned value G is from your objective function. I suspect that you are returning a vector, not a scalar.

Alan Weiss

MATLAB mathematical toolbox documentation