Centralized and Distributed Deep Reinforcement Learning Methods for Downlink Sum-Rate Optimization

For a multi-cell, multi-user, cellular network downlink sum-rate maximization through power allocation is a nonconvex and NP-hard optimization problem. In this article, we present an effective approach to solving this problem through single- and multi-agent actor-critic deep reinforcement learning (DRL). Specifically, we use finite-horizon trust region optimization. Through extensive simulations, we show that we can simultaneously achieve higher spectral efficiency than state-of-the-art optimization algorithms like weighted minimum mean-squared error (WMMSE) and fractional programming (FP), while offering execution times more than two orders of magnitude faster than these approaches. Additionally, the proposed trust region methods demonstrate superior performance and convergence properties than the Advantage Actor-Critic (A2C) DRL algorithm. In contrast to prior approaches, the proposed decentralized DRL approaches allow for distributed optimization with limited CSI and controllable information exchange between BSs while offering competitive performance and reduced training times.