Nonlinear Least Squares implementation

Question

Federico Maglione 2020-1-27

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/502199-nonlinear-least-squares-implementation

评论： Federico Maglione 2020-1-28

Dear all,

I believe my problem is trivial but I do not know I to make it work. I would appreciate someone's help about it. If I define my function myLS like this

function G = myLS(x)
F = [10 30 30];
S0 = 32.2337;
r = 0;
w = 0;
t = [1 5 10];
Q0 = [0.9917 0.8226 0.8081];
fun1 = myequity(x(1), F, t, r, w, x(2)) - S0;
fun2 = myQ(x(1), F, t, r, w, x(2)) - Q0;
G = [fun1 fun2];
end

And then I use the non linear least square

[x,resnorm,res,eflag,output1] = lsqnonlin(@myLS,x0)

The programme finds correctly the values of x which solve the problem. However, I would like to have something like

function G = myLS(F, S0, r, w, t, Q0, x)
fun1 = myequity(x(1), F, t, r, w, x(2)) - S0;
fun2 = myQ(x(1), F, t, r, w, x(2)) - Q0;
G = [fun1 fun2];
end

and solve the same non linear least square problem. It does not work however and the following error appears

Not enough input arguments.
Error in myLS (line 5)
fun1 = myequity(x(1), F, t, r, w, x(2)) - S0;
Error in lsqnonlin (line 205)
            initVals.F = feval(funfcn{3},xCurrent,varargin{:});
Caused by:
    Failure in initial objective function evaluation. LSQNONLIN cannot continue.