Optimizing Power Control in Cellular and Cell-Free Massive MIMO Systems: A SVM/RBF Approach

This paper explores the optimization of power control in both cellular (CL) and cell-free (CF) massive MIMO (mMIMO) systems using a hybrid approach combining support vector machine (SVM) and radial basis function (RBF). The traditional WMMSE method, while effective, exhibits high computational complexity and suboptimal convergence in large-scale systems. The proposed SVM/RBF method addresses these challenges by significantly reducing the computational overhead, as detailed in the computational complexity analysis in Proposed SVM/RBF Methods. To address these challenges, we propose an SVM/RBF-based method for power control (PC) that leverages SVM regression to predict optimal PC vectors and utilizes RBF kernels to enhance prediction accuracy by transforming input features into higher-dimensional spaces. The proposed method dynamically adjusts transmission power levels of user devices based on real-time channel conditions, thereby optimizing resource utilization and system performance. Simulation results demonstrate that the SVM/RBF approach significantly outperforms the WMMSE method in both spectral efficiency and computational efficiency. In terms of Area Under the Curve (AUC) metric, the SVM-RBF method shows a substantial performance gain with AUC values of 24,931 for CL-mMIMO systems compared to 12,698 for WMMSE. Additionally, the SVM-RBF method reduces execution time by approximately 30% in both CL and CF-mMIMO scenarios. This paper confirms that the SVM/RBF method offers a robust, efficient, and scalable solution for optimizing PC in complex wireless communication environments.