AI-Driven Resource Allocation for Energy-Efficient 6G Massive MIMO Networks
Keywords:
6G Wireless Networks, Massive MIMO, Energy Efficiency, Deep Reinforcement Learning (DRL), Proximal Policy Optimization (PPO), Resource Allocation, Power Control, User Scheduling, AI for Wireless Communications, Markov Decision Process (MDP), Spectral Efficiency, Smart Antenna Systems, Next-Generation Wireless Networks, Intelligent Radio Resource Management.Abstract
The fast development of the sixth-generation (6G) wireless networks requires novelty to meet the dual objective of extreme data rate and low-energy requirements. Spectral efficiency Spectral efficiency increases significantly with Massive MIMO, a core 6G technology, since it provides spatial multiplexing. The resource management and power drawbacks of its huge deployment of antennas, however, are problematic. The paper suggests a Deep Reinforcement Learning (DRL)-powered resource allocation architecture that intends to optimize energy efficiency in massive MIMO networks. In particular, the issue is broken down as a markov decision process (MDP) and a proximal policy optimization (PPO) agent is designed to dynamically change the transmission power and schedule the users according to current state of channel and traffic information. The given method learns alternative policies and jointly maximizes throughput and energy minimization with time. The simulation outcomes in a simulated 6G setting with base stations outfitted with 128-antennas and 20 users show that the DRL-based system can save up to 25 percent of total energy consumption, comparing to traditional heuristic-based systems, and have comparable spectral efficiency. In addition, the PPO agent has steady convergence and flexibility to the different traffic needs. These results denote the possibility of smart control measures to resolve the energy-performance trade-off dilemma and suggest the feasibility of DRL toward scalable resource management in 6G and subsequent deployments of massive MIMO.
