Interference-Constrained Pricing for D2D Networks

Yuan Liu (刘元), Yuan Liu, Rui Wang, Zhu HanIEEE Trans. Wireless Communications 2017 · DOI

TL;DR

This paper tackles the challenge of interference management in device-to-device (D2D) networks operating under cellular infrastructure. They propose a novel pricing framework where the base station (BS) charges D2D users for the interference they cause, incentivizing them to manage their power and minimize cross-tier interference. This approach leads to three distinct pricing algorithms (uniform, optimal differentiated, and suboptimal differentiated) that balance the conflicting objectives of the cellular and D2D tiers.

What Problem Does This Paper Address?

Device-to-device (D2D) communication, where devices communicate directly without going through a base station (BS), offers significant benefits like improved spectral efficiency, extended coverage, and offloaded traffic in cellular networks. However, a major hurdle is managing the interference that D2D communications introduce to the existing cellular infrastructure and its users. This cross-tier interference can degrade the performance of cellular services. Existing solutions often focus on resource allocation without explicitly addressing the economic incentives for D2D users to control their interference. This paper aims to bridge this gap by proposing a pricing mechanism, managed by the BS, to control D2D interference while maximizing the BS's revenue.

Methods

The core of the paper's approach is a Stackelberg game formulation, where the BS acts as the leader and D2D users are followers.

  • Interference Pricing: The BS sets prices (π) for the interference caused by each D2D user. D2D users then decide their transmission power to maximize their individual utility, which is their achievable rate minus the price they pay.
  • Noncooperative Power Control Subgame: The D2D users' power allocation decisions are modeled as a noncooperative game. An iterative, distributed power allocation algorithm is proposed to find the Nash Equilibrium (NE) for this subgame. This algorithm is proven to converge to a unique NE.
  • BS Pricing Algorithms:
    • Uniform Pricing: The BS broadcasts a single price to all D2D users. This algorithm is designed for limited network information and is implemented in a fully distributed manner.
  • Optimal Differentiated Pricing: The BS charges different prices to different D2D users based on their individual interference levels. This requires global network information and is partially distributed.
  • Suboptimal Differentiated Pricing: A simplified version of differentiated pricing that offers a closed-form solution, reducing computational complexity and allowing for fully distributed implementation. This relies on assumptions about path loss and interference.

Datasets

Not discussed in the paper. The paper uses simulated data based on random channel realizations and user distributions.

Key Insights and Ideas

  • Interference as a Divisible Resource: The paper frames the BS's interference tolerance margin as a resource that can be "sold" to D2D users through pricing. This economic perspective is key to incentivizing interference management.
  • Stackelberg Game for BS-User Interaction: Modeling the BS as a leader and D2D users as followers effectively captures the hierarchical control and conflicting objectives in D2D-underlay cellular networks.
  • Distributed Power Control for D2D Users: The proposed iterative algorithm for D2D users' power allocation is crucial because it allows for decentralized operation, aligning with the self-organizing nature of D2D networks.
  • Trade-off between BS Revenue and D2D Utility: The simulations clearly demonstrate the inherent conflict: higher prices for interference lead to greater BS revenue but lower achievable rates (and thus utility) for D2D users.
  • Algorithm Efficiency: The paper presents algorithms that are either fully distributed or partially distributed, with some offering closed-form solutions, making them practical for real-world deployment.

Key Results

  • Convergence of Power Allocation: Algorithm 1 for D2D power allocation is shown to converge quickly (within a few iterations) to a unique Nash Equilibrium, regardless of initial conditions, even with a large number of users (up to 100).
  • Performance of Pricing Schemes:
    • Sum Rates: The uniform pricing scheme generally achieves higher sum rates for D2D users, especially at higher Signal-to-Noise Ratios (SNRs).
  • Revenue: Differentiated pricing schemes (both optimal and suboptimal) consistently yield higher revenue for the BS compared to uniform pricing. The optimal differentiated scheme generally performs best in terms of revenue.
  • Impact of Price and Interference Constraint: Simulations show that as the price increases, D2D users reduce their transmit power, leading to lower aggregate interference at the BS. Similarly, a tighter interference temperature constraint (lower $I_{th}$) also influences the pricing and power allocation outcomes.

Limitations and Future Work

  • Myopic Users: The current model assumes D2D users are "myopic," meaning they only consider immediate payoffs. Future work could explore models with foresightful users who optimize for long-term gains.
  • Cooperative Network: The paper assumes users and the BS cooperate by following the proposed algorithms. Future research could investigate how to incentivize truthful reporting of channel state information (CSI) and penalize cheating behaviors.
  • Simplified Channel Models: While Rayleigh fading is used, more complex channel models could be explored.
  • BS Information Requirements: The optimal differentiated pricing requires global network information, which might be challenging to obtain in practice. The suboptimal scheme addresses this by making simplifying assumptions.
  • General Utility Functions: The paper uses achievable rates as utility functions for simplicity. Extending the framework to more general differentiable, increasing, and concave utility functions is a potential future direction.

Summary updated

BibTeX
@article{2017interference,
  author = {Yuan Liu (刘元) and Yuan Liu and Rui Wang and Zhu Han},
  title = {{Interference-Constrained Pricing for D2D Networks}},
  journal = {IEEE Trans. Wireless Communications},
  year = {2017},
  doi = {10.1109/twc.2016.2625255},
}
Figure