Only scalar quadratic forms can be specified in CVX_?

1

Hello to everyone.
how can i solve this problem , Is this a problem because of CVX or it may be main problem is non-convwex?
Do you have a way? please help me
Thank you

for i=1:K
constraint =[];

R = p(i)*alfa(i);
S = alfa(i)*(sum(p)-p(i))

cvx_begin
variables W(1,N)

         constraint = [constraint log(((R/S)+1)-((R/S)*sigma2)*inv_pos(real(trace(H(i,:)'*W*W'*H(i,:)))+sigma2))];
     


maximize(constraint)
subject to

constraint >= QOS

cvx_end

end

2

Someone will most likely ask. Do you have proof that your problem is convex?

Finally showing log output with error messages is recommended.
That way you help us to help you.

3

Not sure actually. But I did similar things with CVX.
What could be the reason for such an error?

Error using * (line 126)
Disciplined convex programming error:
Only scalar quadratic forms can be specified in CVX
.

Error in ploblem_W (line 38)
constraint = [constraint log(((R/S)+1)-((R/S)*sigma2)*inv_pos(real(trace(H(i,:)‘WW’*H(i,:)))+sigma2))];

4

You haven’t followed CVX’s rules. So you need to prove your problem is convex.