Abstract
This paper introduces the weighted-sum energy efficiency (WSEE) as an advanced performance metric designed to represent the uplink energy efficiency (EE) of individual user equipment (UE) in a user-centric Cell-Free massive MIMO (CF-mMIMO) system more accurately. In a realistic user-centric CF-mMIMO context, each UE may exhibit distinct characteristics, such as maximum transmit power limits or specific minimum data rate requirements. By computing the EE of each UE independently and adjusting the weights accordingly, the system can accommodate these unique attributes, thus promoting energy-efficient operation. The uplink WSEE is formulated as a multiple-ratio fractional programming (FP) problem, representing a weighted sum of the EE of individual UEs, which depends on each UE's transmit power and the combining vector at the central processing unit (CPU). To effectively maximize WSEE, we develop optimization algorithms based on the quadratic transform (QT), which is effective for multiple-ratio FP. By applying QT sequentially to each user's EE and the uplink SINR, the method converts the nonconvex WSEE objective into tractable subproblems and ensures stable, monotone convergence. We further introduce an approximate variant that alleviates QT's inherent nonlinearities to accelerate convergence. Compared with global energy efficiency (GEE)-oriented baselines, the proposed algorithms yield simultaneous improvements in user power consumption and spectral efficiency, while also reducing optimization time. Overall, the framework provides a foundation for designing operational strategies tailored to specific system requirements.
Introduction
As wireless networks evolve toward 6G, Energy Efficiency (EE) has become a primary design objective alongside Spectral Efficiency (SE). Cell-Free Massive MIMO (CF-mMIMO) systems, where distributed Access Points (APs) cooperatively serve User Equipment (UE), offer a promising architecture to meet these goals. However, traditional metrics like Global Energy Efficiency (GEE)—the ratio of total network data rate to total power consumption—often prioritize aggregate performance at the expense of individual users. This paper addresses this limitation by introducing Weighted-Sum Energy Efficiency (WSEE) as a more equitable performance metric for user-centric CF-mMIMO systems.
The Challenge of Fairness in Cell-Free Networks
Standard CF-mMIMO research often focuses on maximizing GEE. While GEE provides a high-level view of network efficiency, it is inherently "unfair" in the uplink phase. GEE optimization may significantly degrade the performance of specific UEs to achieve a marginal gain in the total average, potentially leading to rapid battery depletion for disadvantaged users or failure to meet their minimum Quality of Service (QoS) requirements.
To solve this, the authors propose the WSEE metric. Unlike GEE, WSEE calculates the EE of each UE independently and applies priority weights. This allows network operators to tailor performance based on specific UE characteristics, such as maximum transmit power limits or high-priority data needs, ensuring that no user is "starved" of resources in the pursuit of global averages.
Optimization via Nested Quadratic Transform
Maximizing WSEE is a formidable mathematical challenge. The objective function is a multiple-ratio fractional programming (FP) problem, which is non-concave and notoriously difficult to solve globally. The complexity is compounded in user-centric systems where the combining vector at the Central Processing Unit (CPU) and the transmit power of each UE must be jointly optimized.
The authors evaluate two main approaches:
- Dinkelbach-like Transform: While effective for single-ratio problems, this classical method fails to guarantee monotonic convergence in multiple-ratio scenarios like WSEE, often leading to fluctuating objective values and suboptimal solutions.
- Nested Quadratic Transform (QT): This is the paper's core contribution. The WSEE objective contains "nested" ratios: the energy efficiency ratio itself contains the Signal-to-Interference-plus-Noise Ratio (SINR). By applying the Quadratic Transform twice in a nested fashion, the authors convert the non-convex problem into a sequence of tractable, concave subproblems.
This nested approach theoretically guarantees convergence to a local optimum, providing a stable and reliable framework for network optimization that avoids the instabilities seen in Dinkelbach-like methods.
Key Insights and Performance Results
The proposed Nested-QT algorithm was tested in a simulated environment featuring 256 APs and 16 UEs. The results highlight several key advantages:
- Efficiency Gains: Nested-QT achieved an 18% improvement in WSEE compared to a baseline of full-power transmission and outperformed the Dinkelbach-like scheme by 10% at convergence.
- Effective Prioritization: By adjusting weights, the system successfully prioritized high-demand users. For instance, giving high-priority UEs ten times the weight of others allowed the system to maximize their EE while still maintaining the minimum required data rates for low-priority users.
- Equitable Distribution: Cumulative Distribution Function (CDF) analysis showed that WSEE optimization provides a much fairer distribution of energy efficiency across the user population compared to GEE-based methods, which tend to starve users with poor channel conditions.
Conclusion and Future Directions
The introduction of WSEE and the Nested-QT optimization framework provides a robust foundation for energy-efficient operation in next-generation CF-mMIMO networks. While the Nested-QT algorithm ensures convergence and high performance, it currently requires more iterations than simpler, less reliable methods. Future research will focus on accelerating this convergence and developing closed-form solutions that reduce reliance on external convex solvers, making the system more suitable for real-time, large-scale deployment.
Summary updated