The optimization problem and my code is shown above. Can I ask if there is something wrong with the writing of this code and why the result solved does not satisfy the inequality constraint above the red line?
As I in indicated in my post at 请问这个SDP问题里面的log项如何进行表达 - #3 by Mark_L_Stone, I don’t understand what the red underlined inequality constraint is defined to be.
You haven’t shown us the output from running CVX or the evidence you have that the constraint is not satisfied, and by how much it is not satisfied (if violated by < 1e-6, that may just be a solver tolerance issue. Is the constraint supposed to be satisfied on each SCA iteration, or only upon convergence of SCA. I have no reason to think that SCA will converge for every set of input data and starting value, or that it converged in your case unless I see evidence that it did.
Here are some comments on SCA: The result obtained using the SeDuMi solver is inaccurate. The optimal solution is not as good as before the optimization. - #12 by Mark_L_Stone