Abstract
This paper investigates pilot and data power optimization for cell-free massive MIMO (CF-mMIMO). We propose an iterative algorithm that jointly updates pilot and data power levels to improve channel estimation and ensure reliable data transmission. Pilot powers are allocated based on the normalized mean square error (NMSE) of channel estimation, granting higher power to users with poor estimates while reducing interference for users with favorable conditions. Based on the resulting channel state information (CSI), data powers are then optimized via geometric programming to achieve max–min fairness across users. By alternating between NMSE-driven pilot power control and fairness-oriented data power allocation until convergence, the proposed method achieves efficient CSI acquisition, balanced interference management, and enhanced fairness. In addition, we introduce a lightweight access point (AP)–user association algorithm that ranks AP–user channel strengths, limits the number of users per AP, and employs iterative replacement to ensure scalability and full user connectivity. Simulation results demonstrate that the proposed framework significantly improves spectral efficiency and fairness compared to conventional methods, while remaining suitable for practical CF-mMIMO deployments.