Utility Maximization for Large-Scale Cell-Free Massive MIMO Downlink

We consider utility maximization problems in the downlink cell-free massive multiple-input multiple-output (MIMO) whereby a large number of access points (APs) simultaneously serve group users. Four fundamental objectives are interest: (i) average spectral efficiency (SE), (ii) proportional fairness, (iii) harmonic-rate, and (iv) minimum SE all users, subject to sum power constraint at each AP. As considered non-convex, existing solutions normally rely on successive convex approximation (SCA) use off-the-shelf solvers, which implement an interior-point algorithm, solve derived problems. The complexity such methods scales quickly with problem size. Therefore, we propose accelerated projected gradient method Particularly, iteration proposed solution is given closed form only requires first order oracle objective, rather than Hessian matrix as known solutions, thus much more memory efficient. Numerical results demonstrate that our achieves same performance but far less run-time, compared SCA method. Simulation show large-scale MIMO has intrinsic user i.e. four functions can deliver nearly uniformed services