Abstract
In a coordinated multipoint (CoMP) transmission system with centralized architecture, minimizing total power consumption while adhering to backhaul capacity and minimum target data rate constraints is a critical challenge. This problem is formulated as a complex mixed-integer optimization. To address this, we developed a joint user selection and rate adaptation scheme (WFR-PS) based on water-filling rate adaptation and a power-saving criterion. Numerical results demonstrate that the proposed algorithm significantly reduces power consumption compared to norm-based and semi-orthogonal user selection (SUS) algorithms, achieving near-optimal performance with reduced computational complexity.
Introduction
Coordinated Multipoint (CoMP) transmission is a key technology for enhancing "cell-edge" performance by allowing multiple base stations to cooperate. However, this cooperation introduces two major hurdles: the backhaul bottleneck and high power consumption. While most research focuses on maximizing total throughput (sum rate), this paper addresses the inverse problem: meeting specific data demands with minimal power. The coupling between user selection and rate adaptation makes this a difficult mixed-integer optimization problem.
The WFR-PS Algorithm
The proposed WFR-PS (Water-Filling Rate adaptation and Power Saving criterion) scheme employs a two-stage heuristic approach to solve the power minimization problem efficiently.
- Zero-Forcing Beamforming (ZFBF): Used to eliminate inter-user interference within a cluster.
- Water-Filling Rate (WFR) Adaptation: Unlike traditional power-based water-filling, this adapts data rates for a fixed user set to minimize power while saturating the backhaul capacity.
- Power Saving (PS) Criterion: An iterative, greedy process that adds users only if they reduce the total required power.
- Search Scope Shrinking: Heuristically removes unlikely candidate users to accelerate the search.
Performance Evaluation
The scheme was evaluated using Monte Carlo simulations modeling a 3-base station cluster with 8 antennas each and up to 100 candidate users. The simulations utilized log-distance path loss and Rayleigh fading across 1000 independent channel realizations.
Key Results
The WFR-PS algorithm demonstrates substantial improvements in power efficiency. As shown in the performance comparison, the proposed method achieves nearly the same power reduction as an exhaustive search, significantly outperforming standard norm-based and SUS-based selection methods.
Specifically, the algorithm finds the optimal power-minimizing solution in approximately 98% of cases. This near-optimality is achieved with a fraction of the computational cost required for an exhaustive search, making it suitable for practical implementation in resource-constrained hardware.
Conclusion
The WFR-PS scheme effectively balances the trade-off between performance and complexity. By jointly optimizing user selection and rate adaptation, it ensures that backhaul capacity is fully utilized and target rates are met with minimum power. Future work may address the impact of imperfect channel state information (CSI) and the extension of these heuristics to distributed CoMP architectures.
Summary updated