DRL Challenges in Wireless Networks

Although the DRL techniques have many advantages in dealing with wireless communications and networking problems at the different layers, some technical limitations still need to be resolved. This chapter will review some challenges in implementing DRL approaches in wireless networks in practice, e.g. the curse of dimensionality, online learning, offline training processes, and quick adaptation to the uncertainty and dynamics of wireless environments. Alternatively, in wireless networks, several open issues in deploying DRL, e.g. inheriting and transferring knowledge among intelligent network entities, will be discussed.