The assignment cannot be performed because the size of the left side is 1×1 and the size of the right side is 0×0.

The object function I code is as below, when I use the ga to call this function , occur the error:"The assignment cannot be performed because the size of the left side is 1×1 and the size of the right side is 0×0." But I do not know where is wrong.Anyone who can help me ? The code and the data needed I have upload,please help me. Thanks advanced.

function F = evaluate_objective(x)
global y_e beta lamda sample1 r0
f = zeros(1,3);%
Xt = [x(1) x(2) x(3) x(4) x(5) x(6) x(7) x(8) x(9) x(10)];
B = [1 1+sum(Xt)];% 一次项
[mS,nS]=size(sample1);[mX,nX]=size(Xt);
if nX~=nS %check that both matrices are of the right shape
    Xt=Xt';
    [mX,nX]=size(Xt);
end
Phi = zeros(mX,mS); %compute pairwise distances of points in X and S
for ii = 1:mX
    for jj = 1:mS
        Phi(ii,jj) = exp(-0.5*((norm(Xt(ii,:)-sample1(jj,:)))./r0)^2);
    end 
end
f(1) = abs((y_e(:,1) - (Phi*lamda(:,1)+B*beta(:,1))))/y_e(:,1);
f(2) = abs((y_e(:,2) - (Phi*lamda(:,2)+B*beta(:,2))))/y_e(:,2);
f(3) = abs((y_e(:,3) - (Phi*lamda(:,3)+B*beta(:,3))))/y_e(:,3);
F = max(f);
end

Answer 1

[Stephen23]

"But I do not know where is wrong"

The problem is caused by using GLOBAL variables instead of parameterizing the function properly:

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

One/some/all of those global variables are empty (probably because they are still set to the default empty array) and so the calculations on the RHS of these lines are empty:

f(1) = .. empty

f(2) = .. empty

f(3) = .. empty

But clearly you cannot assign zero elements to one element, thus the error.

The solution, as I stated at the start, is to parameterize the function properly:

S = load('var.mat');
fnh = @(x) evaluate_objective(x, S.Y, S.beta, S.lamda, S.sample1, S.r0);
out = ga(fnh,..)

But becuse your function has other errors, I cannot run this.

[zx yang]

Thanks for your advice. But the error occur in the line

" Phi(ii,jj) = exp(-0.5*((norm(Xt(ii,:)-sample1(jj,:)))./r0)^2);"

so maybe the problem is not caused by the global variables , and I have already assigned the global variables before call this function .

[zx yang]

well，the ‘x’ in the function is the unknown number, which I need to calculate through the Ga .

I still can not found the error. but thanks a lot .

[Stephen23]

"But the error occur in the line ...so maybe the problem is not caused by the global variables"

Exactly the same reasoning applies to that line too: both sample1 and r0 are global (i.e. empty).

"I have already assigned the global variables before call this function"

But as you do not show us the code where you do that assignment I can safely presume that it is buggy.

[zx yang]

"But as you do not show us the code where you do that assignment"

the data used to the assignment is in the var.mat

[Stephen23]

"the data used to the assignment is in the var.mat"

There is no problem with the data.

The problem is the code, which for some reason you do not want to show us. I am happy to help you, but I cannot debug code that I cannot see.

Please show us the code where those global variables are defined.

My guess (and this is just a guess) is that you did not declare them to be global and just expect them to jump into that other workspace. In any case, the solution is still the same: avoid global variables and write more robust, more efficient code by parameterizing the function.

[zx yang]

ok, thanks。I upload my code which I used to do the updating。

and as for the 'Y' var,use that in the mat is enough ,because it refer to huge calculation.

as for the var 'loop' in the code ,it;s only a number you can assign as you want

and the LHS is a kind of sampling you see and you will know

Thanks

[Stephen23]

@zx yang: thank you for uploading your code.

I note that you modified the function signature and removed the global variables... you just need to parameterize the function as I showed in my answer. For example, something like this:

fnh = @(x) evaluate_objective(x, Y, beta, lamda, sample1, r0);
out = ga(fnh,10,[],[],[],[],LB,UB,[],[],options);

[zx yang]

Thank you！！ I have a try.

[zx yang]

Anyway thanks a lot. Thank you again !!