Introduction
As mobile data traffic continues its exponential growth, Device-to-Device (D2D) communication has emerged as a critical technology for offloading traffic from base stations and enhancing spectral efficiency (SE). By allowing devices in close proximity to communicate directly, network capacity can be significantly expanded. However, integrating D2D into existing cellular frameworks presents two primary challenges: energy constraints and spectrum scarcity.
D2D devices are typically battery-constrained. While Energy Harvesting (EH) offers a sustainable solution, the stochastic nature of energy arrival makes transmission scheduling complex. Furthermore, to maximize spectrum utility, D2D links often reuse cellular resource blocks (RBs). Traditional models often restrict this to one-to-one or many-to-one mappings, which may underutilize available resources. This paper proposes a many-to-many matching game to jointly optimize spectrum matching, power allocation, and transmission duration.
System Architecture
The study considers a single-cell network where multiple D2D links coexist with cellular users (CUs). The D2D transmitters are equipped with energy harvesting units, allowing them to scavenge energy from the environment to power their transmissions.
In this underlay scenario, D2D links share the uplink spectrum of CUs. The objective is to maximize the sum spectral efficiency of the D2D tier while ensuring the Quality of Service (QoS) for CUs and adhering to the causal energy constraints of the EH-enabled devices.
Energy Harvesting and Transmission Model
A key innovation of this work is the detailed modeling of the energy harvesting process and its impact on transmission timing. Energy arrives at the D2D transmitters following a Poisson process. Because energy is harvested intermittently, the transmitter must carefully manage its "energy budget" across different time slots.
The optimization problem is formulated as a mixed-integer nonlinear programming (MINLP) problem, which is NP-hard. To solve it efficiently, the authors decouple the problem into spectrum matching and continuous resource allocation (power and time).
The Many-to-Many Matching Game
The spectrum allocation is modeled as a many-to-many matching game between D2D links and cellular RBs. Unlike simpler models, this approach allows a single D2D link to utilize multiple RBs simultaneously, providing a "diversity gain" that improves robustness against interference.
The game incorporates "externalities," recognizing that a D2D link's preference for an RB depends on the interference generated by other links sharing that same resource. The algorithm employs swap operations to reach a "two-sided exchange stable" state, where no pair of D2D links can improve their utility by swapping their assigned RBs.
Results and Key Insights
Numerical simulations demonstrate significant performance gains over traditional allocation strategies:
- Spectral Efficiency: The many-to-many approach outperformed one-to-one matching by 47% and many-to-one matching by 5%.
- Optimality: The proposed algorithm achieves approximately 95% of the performance of an exhaustive search while maintaining much lower computational complexity.
- QoS and EH Robustness: The system effectively protects CU data rates and scales efficiently with increasing energy harvesting rates. As the energy arrival rate increases, the sum SE improves because devices can afford longer transmission durations.
Conclusion
This research highlights the potential of many-to-many matching games in managing complex, interference-limited D2D networks. By accounting for energy harvesting constraints and allowing flexible spectrum reuse, the proposed framework achieves near-optimal spectral efficiency. Future work may extend this model to dynamic environments with mobile nodes or more sophisticated energy arrival patterns.
Summary updated