Now that you showed the whole program, which is different than what you posted before, I see that you are calling CVX iteratively within SCA. Therefore, my first post applies, specifically, including everything in the link I provided.
Perhaps you should consider trying to use a non-convex solver under YALMIP.
If you do stay with CVX, I recommend you use Mosek as solver if available to you, otherwise install CVXQUAD with its ecponential.m replacement, as discussed in CVXQUAD: How to use CVXQUAD's Pade Approximant instead of CVX's unreliable Successive Approximation for GP mode, log, exp, entr, rel_entr, kl_div, log_det, det_rootn, exponential cone. CVXQUAD's Quantum (Matrix) Entropy & Matrix Log related functions
However, it does appear that CVX’s Successive Approximation method did solve your problem and that the difficulty in your case is that SCA is a lousy algorithm, which neither Mosek not CVXQUAD will fix for you.