Failure in initial objective function evaluation. FSOLVE cannot continue.

Recently, i'm trying to solve a nonlinear equation set by using "fsolve" . I created my personal function as the following:

function F = poSTAR_14(x,Hm_RI,at_RI,J)

F(1)=Hm_RI*x;

F(2)=at_RI*x-[1;0];

F(3)=abs(x(1:32,1)).^2+abs(x(33:64,1)).^2-ones(J,1);

end

Above is the function that contains 3 nonlinear equations and I saved it as a .m file with my main file.

Below is the part that I call the fsolve to solve my problem, where , are known vetors, is known scalar, initial point

Hm_R=real(Hm).*100;

Hm_I=imag(Hm).*100;

Hm_RI=[Hm_R,-Hm_I;Hm_R,Hm_I];

at_R=real(as_ele_Tx.');

at_I=imag(as_ele_Tx.');

at_RI=[at_R,-at_I;at_R,at_I];

w_R=real(conj(W_Tx));

w_I=imag(conj(W_Tx));

fun = @poSTAR_14;

x0 = [w_R;w_I];

x = fsolve(@(x)fun(x,Hm_RI,at_RI,J),x0);

But something is wrong with my code and It seems someting wrong with the size of my matrix in the problem formulation:

Unable to perform assignment because the indices on the left side are not compatible with the

size of the right side.

Error in poSTAR_14 (line 2)

F(1)=Hm_RI*x;

Error in fsolve_PO>@(x)fun(x,Hm_RI,at_RI,J) (line 202)

x = fsolve(@(x)fun(x,Hm_RI,at_RI,J),x0);

Error in fsolve (line 248)

fuser = feval(funfcn{3},x,varargin{:});

Error in fsolve_PO (line 202)

x = fsolve(@(x)fun(x,Hm_RI,at_RI,J),x0);

Caused by:

Failure in initial objective function evaluation. FSOLVE cannot continue.

[Jon]

I am trying to reproduce your problems but I need to have an example value to use for Hm

[Jon]

Also please provide example values for as_ele_Tx and W_Tx,

[奥 刘]

@Jon Sorry I'm now away from my PC that run this matlab code, I will provide the data as soon as I get back to work. Really thanks for your help!

[奥 刘]

@Jon Thanks for your help! I have solved this problem.

Answer 1

[Matt J]

function F = poSTAR_14(x,Hm_RI,at_RI,J) 

 F1=Hm_RI*x;
 F2=at_RI*x-[1;0];
 F3=abs(x(1:32,1)).^2+abs(x(33:64,1)).^2-ones(J,1); %The abs() seems unnecessary here
 
 F=[F1(:);F2(:);F3(:)];
end

[奥 刘]

Thanks! It worked. Could you please tell me why my code cannot proceed?

[Matt J]

You're welcome, but please Accept-click the answer to indicate that it provided a solution.

Your code original code won't work because the left hand side F(1) is a scalar location. The right hand side Hm_RI*x is not a scalar, so it cannot fit there, and similarly with the other F(i).

[奥 刘]

@Matt J

After the correction, the command window showed that,

Solver stopped prematurely.

fsolve stopped because it exceeded the function evaluation limit,

options.MaxFunctionEvaluations = 6.400000e+03.

Does it mean I should pick a better starting point or just set a larger "MaxFunctionEvaluations"?

[奥 刘]

@Matt J Thanks! Now I know why I am wrong.