Decentralized power allocation for coordinated multiple access channels

E. Veronica Belmega, Samson Lasaulce, Mérouane Debbah2008 · DOI

Abstract

In this paper, a game theoretic approach is used to derive the optimal power allocation in multiple access channels (MAC) for decentralized systems. Unlike previous results, a simple coordination mechanism is used between selfish users. The nature and influence of the coordination mechanism is studied to determine how decentralized behavior can approach centralized efficiency.

Introduction

In standard wireless Multiple Access Channels (MAC), multiple users transmit data to a single base station. Traditional centralized systems rely on a central controller to dictate power levels and data rates for every user. While highly efficient, this approach incurs significant signaling overhead that struggles to scale with large, dynamic networks.

In contrast, decentralized systems allow "selfish" users to choose their own power levels to maximize individual data rates. Historically, this lack of coordination leads to a "tragedy of the commons," where overall network performance suffers. This research explores the minimum coordination required to align individual selfishness with social efficiency, maximizing total network throughput.

Game Theoretic Coordination

The authors frame the power allocation problem as a non-cooperative game. The core of the solution is a simple coordination mechanism: a low-bandwidth "coin flip" signal from the base station that determines the decoding order for Successive Interference Cancellation (SIC).

Because the user decoded last experiences no interference from others, the decoding order significantly impacts individual rates. By setting a probability ($p$) for a specific decoding order, the base station establishes the "rules of the game." Users then adapt their power strategies to maximize their average rate given this probability. The authors prove the existence and uniqueness of a Nash Equilibrium (NE) in this setting, where no user can improve their rate by unilaterally changing their strategy.

Achievable rate region for a 2-user MIMO MAC system ($n_t=n_r=4$, $\rho=5$ dB) comparing SIC decoding to the centralized sum rate.

For complex multi-antenna (MIMO) systems, the research utilizes "large system" limits. By assuming the number of antennas approaches infinity, complex equations simplify into predictable approximations that reveal the underlying physics of the system.

Results and Efficiency

The findings demonstrate that simple coordination is remarkably powerful. In single-antenna (SISO) systems, decentralized games achieve 100% of the centralized sum-capacity regardless of the coordination signal's distribution. In these cases, selfishness and social efficiency are perfectly aligned.

In multi-antenna fading channels, a performance gap emerges, but it is surprisingly small. Simulations of MIMO systems show that the total network rate is a stable function of the coordination probability $p$. Even in the worst-case coordination scenario, the performance remains within approximately 1% of the centralized optimum.

The achievable sum rate $R_1+R_2$ vs. coordination probability $p$ for a MIMO MAC system ($n_t=n_r=4$, $\rho=5$ dB).

This stability is attributed to a "double averaging" effect in large MIMO systems, where time-averaging (fading) and spatial-averaging (multiple antennas) create a highly predictable environment. This allows the base station to "slide" performance along the optimal capacity boundary by simply adjusting $p$, favoring specific users without compromising total network speed.

Conclusion and Future Work

This research proves that decentralized wireless networks do not require heavy-handed centralized control to be efficient. A simple coordination signal is sufficient to drive selfish users toward near-optimal performance.

Future work may address current limitations, such as the assumption of perfect channel state information at the receiver (CSIR) and the focus on 2-user scenarios. As the number of users ($K$) grows, the complexity of the coordination signal ($K!$ possible orders) may present new challenges. Additionally, exploring how these rules adapt to dynamic environments where users frequently enter and leave the network remains a vital area for further study.

Summary updated

Abstract

In this paper, a game theoretic approach is used to derive the optimal power allocation in multiple access channels (MAC) for decentralized systems. Unlike previous results, a simple coordination mechanism is used between selfish users. The nature and influence of the coordination mechanism is stu
BibTeX
@misc{belmega2008decentralized,
  author = {E. Veronica Belmega and Samson Lasaulce and Mérouane Debbah},
  title = {{Decentralized power allocation for coordinated multiple access channels}},
  year = {2008},
  doi = {10.4108/icst.wiopt2008.3213},
}
Figure