I am trying to solve the following problem using CVX. Indeed, someone has claimed that it is solved using CVX already and I would like to redo what has been done before:

maximize \sum\sum

when I attempt the following code, I always get “Invalid quadratic form(s): not a square.” error and I am pretty stuck using CVX to deal with this problem.

r = [54 54 54;
54 54 54;
54 54 54;
54 54 54;
54 54 54];
[m,n] = size®; % m X Users & n X APs

cvx_begin
variable x(m,n)
variable b(m,n)
maximize sum(sum(x(1:m,1:n).*b(1:m,1:n) ))
subject to
sum(x,2) == 1
b == r ./ sum(x,1)
0 <= x <= 1
cvx_end

Regarding to the author, x can be considered as a real variable for relaxation and after solving the problem by using rounding technique we can solve the original problem. I really appreciate your feedback. thanks.

I have changed the format of the program as follows and I used geometric mode, but still I get error to compile the objective function! any feedback? (P.S.: I ignored log function at the moment)

Oh, I agree, it’s not professional. I wasn’t trying to be. I would suggest you recognize this forum is provided and paid for as a public service, and run on a volunteer basis. If we don’t have an answer for you, that’s just the way it is.