Energy efficient RF chains selection based on integrated Hungarian and genetic approaches for uplink cell-free millimetre-wave massive MIMO systems

The purpose of this work is to explore the decrease of total used power in cell-free millimetre-wave (mm-Wave) massive multiple-input multiple-output (MIMO) systems, which can be regarded an essential technology for future wireless generations to improve system performance. One of the most important strategies for reducing total power consumption is to activate and deactivate radio frequency (RF) chains at each access point (AP) in the coverage region. Nonetheless, the optimization issue for this methodology is NP-hard, and an exhaustive search method may be used to determine the ideal number of RF chains at each AP in the cell-free network. Unfortunately, the exhaustive searching approach is prohibitively complicated, indicating that it is unworkable when there are a significant number of APs in the service region. Furthermore, present RF chain selection approaches prioritize decreasing consumed power and complexity at the price of system performance in terms of total possible rate. This research solves this issue by introducing a unique RF chain selection approach that combines Hungarian and genetic algorithms. The fact that the genetic algorithm (GA) can readily determine the ideal number of active RF chains, yet this technique generally entails significant complexity in large-scale cell-free networks, prompted this notion. As a result, the Hungarian method is used early in the GA to overcome the complexity problem while still retaining system performance. In addition, the suggested system employs a semi-centralized hybrid beamforming architecture in which all analogue combiners for all APs are operated at a central processing unit using channel state information. In addition, each AP has a fully linked phase shifters network and restricted RF chains connecting to its antennas. Finally, simulation findings reveal that, when compared to state-of-the-art techniques, the suggested approach achieves the maximum attainable rate and overall energy efficiency with a tolerable computational complexity.