Error using fminsearch and optimset

Hi,

I am trying to calibrate parameters using experimental data and simulation results from simulink. But when I try to run it, Matlab gives me some errors. Below the errors:

>> CalibrationASM1_Namoniacal

Index exceeds matrix dimensions.

Error in optASM1 (line 10)

bh=p(3);

Error in fminsearch (line 189)

fv(:,1) = funfcn(x,varargin{:});

Error in CalibrationASM1_Namoniacal (line 11)

[p,fval,exitflag,output]=fminsearch(@optASM1,[1,1],options)

Below the way I put my equation and inputs in matlab:

% Load parameters for ASM1 first
%% Variable input
t=xlsread('Resultados experimentais.xlsx','N amoniacal','l7:l16');
Dexp=xlsread('Resultados experimentais.xlsx','N amoniacal','m7:m16');
global yh;
global ya;
global bh;
global ba;
%% Optimization
options=optimset('MaxFunEvals',400,'TolFun',1e-8,'TolX',1e-6);
[p,fval,exitflag,output]=fminsearch(@optASM1,[1,1],options)
%% Outputs
sim('ASM1_batelada');
Dsim=Snhout([2 25 41 49 67 77 83 86 90 92])
yh
ya
bh
ba
cc=corrcoef(Dexp,Dsim);
r2=cc(2,1)

And my functions is written in another file optASM1, as follows:

function f=optASM1(p)
Dexp=xlsread('Resultados experimentais.xlsx','N amoniacal','l7:l16');
texp=xlsread('Resultados experimentais.xlsx','N amoniacal','m7:m16');
global yh;
global ya;
global bh;
global ba;
yh=p(1);
ya=p(2);
bh=p(3);
ba=p(4);
%% Simulink
sim('ASM1_batelada');
Dsim=Snhout([2 25 41 49 67 77 83 86 90 92])
%% Cost-function
f=sum((Dexp-Dsim).^2)

Please, need help to solve this problem.

Answer 1

[Walter Roberson]

You are passing [1,1] as the second argument to fminsearch(). That becomes the input to the first call to optASM1, where it becomes the p argument. You attempt to access p(3) and p(4) but that input only has two values. If you access 4 values in your function then your initial input needs to be length 4 or more.

We recomment that you do not use global; it is often difficult to debug problems where global variables are used.

http://www.mathworks.com/help/matlab/math/parameterizing-functions.html

[Nara Salles]

Hi Walter, thanks a lot for your help. The errors are not happening anymore but one thing is that the calibration is running for 14 hours and it hasn't finished yet. Do you know if it is normal or if there is something that I can do to reduce the calibration time?.

Thank you again for your attention. Nara

[Walter Roberson]

Could you post your current code, and the xlsx file?

[Nara Salles]

Hi Walter, I have stopped the optimazation and changed the .m file.

Removing global as you recommended, it is working perfectly.

I am very thankful!!. Best regards, Nara.