Optimisation toolbox error

I have the following function which I wan to use in optimisation toolbox:

function diff = fit_simp_optim()

% This function is called by lsqnonlin.
% x is a vector which contains the coefficients of the
% equation.  X and Y are the option data sets that were
% passed to lsqnonlin.
X=xlsread('QS2T.xls',9,'V638:C1835');
Y=xlsread('QS2T.xls',9,'W638:W1835');

diff = (A+B*(X.^n)) - Y;

end

Now, in my command window I type:

>>optimtool

I get the optimisation toolbox;

In the optimisation toolbox,

1)corresponding to to the solver i select, lsqnonlin

2)then corresponding to algorithm, I select Trust region reflective

3)then corresponding to objective function i type @fit_simp_optim() which is the name of my .m function file

4)derivatives approximated by solver

5)Start point i give (which is for constants A,B,C- see my function above):[800;1537;0.1448]

6)Lower bound as [800;0;0] and Upper bound as [1600;6000;10]

Howver, when i start, it starts optimising but then says 'too many input arguments'

Please can anyoone help.

Answer 1

[Walter Roberson]

The function you pass to lsqnonlin must accept a vector of values the same size as your starting point, and must return a vector of function values (possibly a different size.)

Have another look at your fit_simp_optim code: you use A and B and n but those are not defined as variables anywhere. (What happened to C, by the way?)

Redefine your code:

function diff = fit_simp_optim(ABn, X, Y)

    % This function is called by lsqnonlin.
    % x is a vector which contains the coefficients of the
    % equation.  X and Y are the option data sets that were
    % passed to lsqnonlin.

    A = ABC(1); B = ABC(2); n = ABC(3);

    diff = (A+B*(X.^n)) - Y;

end

Then code

X = xlsread('QS2T.xls',9,'V638:C1835');
Y = xlsread('QS2T.xls',9,'W638:W1835');
lb = [800;0;0];
ub = [1600;600;10];
StartAt = [800;1537;0.1448];
optABn = lsqnonlin( @(ABn) fit_simp_optim(ABn,X,Y), StartAt, lb, ub);

[Shalini]

Thanks a lot..it works.Can you tell me how to see the R-square for the fitted curve?

[Walter Roberson]

I am not certain, but it looks to me as this might be the second output argument:

[x,resnorm] = lsqnonlin(...) returns the value of the squared 2-norm of the residual at x: sum(fun(x).^2).

If so, then

[optABn, R2] = lsqnonlin( @(ABn) fit_simp_optim(ABn,X,Y), StartAt, lb, ub);