Maximizing the Energy Efficiency in Integrated Sub-6 GHz, mmWave and THz Wireless Networks

To cope with the spectrum scarcity of emerging mobile applications, using higher frequency bands (such as THz) becomes a necessity. Combining simultaneously higher and lower bands is then an important step for future cellular networks. Meanwhile, resource management plays a dominant effect on the system performance especially when different quality of service requirements are considered. In this paper, we formulate and investigate a joint optimization problem of resource allocation in integrated sub-6 GHz, mmWave and THz networks to maximize the system energy efficiency (EE). Therefore, we propose efficient centralized and distributed low-complexity greedy solutions. Also, we propose more efficient multi-agent reinforcement learning (MARL) solutions where users, modeled as agents, collaborate to learn and converge to the optimal user association that maximizes the EE. Simulation results show the EE provided by the proposed solutions and illustrate the superiority of the MARL-based solutions.