Antenna Selection of Cell-Free XL-MIMO Systems with Multi-Agent Reinforcement Learning

Cell-free massive multiple-input multiple-output (mMIMO) and extremely large-scale MIMO (XL-MIMO) have been investigated as promising innovations for next-generation wireless communication systems. In this article, we achieve the combination of the aforementioned technologies and explore such an extended paradigm namely cell-free XL-MIMO. In such systems, high-dimensional matrix operations due to massive antennas consume considerable computational resources. Moreover, with the significant benefits of increasing the number of degrees of freedom, it is necessary to simultaneously reduce power consumption. Therefore, we aim to achieve the antenna selection strategy in cell-free XL-MIMO systems with multi-agent reinforcement learning (MARL). And we investigate the key performance indicators of cell-free XL-MIMO systems, such as spectral efficiency (SE) and energy efficiency (EE). The numerical results and analyses show a significant EE increase and uniform SE coverage in cell-free XL-MIMO systems with MARL methods.