Power Minimization in Multi-pair Two-Way Relaying

Xiaojun YuanarXiv 2013 · arXiv

TL;DR

This paper bridges the gap between two common problem formulations in wireless relay networks: minimizing total transmit power and maximizing the minimum signal-to-interference-plus-noise ratio (SINR). The authors demonstrate that these seemingly different problems are, in fact, equivalent under certain conditions, offering flexibility in choosing the most computationally tractable approach for designing optimal relay beamforming strategies.

What Problem Does This Paper Address?

In multi-pair two-way relay networks, a critical challenge is to design the relay's transmit beamforming to efficiently manage power while ensuring reliable communication for all users. Two standard approaches exist: minimizing the total transmit power subject to SINR constraints, and maximizing the minimum SINR across all users. While both aim for efficient and fair communication, they are often treated as distinct optimization problems. This paper addresses the question of how these two formulations relate to each other and whether one can be used to solve the other, thereby simplifying the design process.

Methods

The core of this paper lies in establishing a mathematical equivalence between two optimization problems.

  1. Power Minimization Problem: This problem aims to minimize the total transmit power of the relay subject to achieving a minimum SINR for each communication link. The authors present this as a convex optimization problem that can be efficiently solved using Semidefinite Programming (SDP).
  2. Max-Min SINR Problem: This problem, referenced from prior work ([1]), aims to maximize the lowest SINR experienced by any user, given a total power budget.

The authors prove that the optimal solution to the power minimization problem can be obtained by solving the max-min SINR problem with a specific power budget. Conversely, they show that the optimal solution to the max-min SINR problem can be found by solving the power minimization problem with a power budget that results in a specific SINR target. This equivalence is established through a series of lemmas that demonstrate the monotonic relationship between power and SINR, and how solutions from one problem can be mapped to the other.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Equivalence of Power Minimization and Max-Min SINR: The most significant insight is the demonstration that minimizing total power under SINR constraints is mathematically equivalent to maximizing the minimum SINR under a power budget, provided the correct parameters are chosen.
  • Flexibility in Problem Formulation: This equivalence offers significant practical advantages. Researchers and engineers can now choose the problem formulation that is computationally easier or more convenient for their specific design goals, without sacrificing optimality.
  • SDP and Bisection Search Synergy: The paper highlights how existing efficient solvers (SDP for power minimization) and standard numerical techniques (bisection search for finding the correct power budget) can be combined to solve both problem types.

Key Results

The paper establishes that:

  • The power minimization problem (2) can be efficiently solved using a single SDP.
  • The optimal solution to the power minimization problem (2) is the same as the optimal solution to the max-min SINR problem (8) in [1] when the power budget $P_R$ is set such that the max-min SINR is exactly 1.
  • The max-min SINR problem (5) in [1] can be solved by finding a parameter $\lambda$ that satisfies a specific equation, which can be achieved using a one-dimensional bisection search over $\lambda$.

Limitations and Future Work

  • Assumptions: The derivations rely on the assumption that the SINR is a strictly increasing function of the power budget, which is generally true in wireless systems.
  • Future Work: The paper concludes by stating that this equivalence allows for the selection of the most tractable form to pursue optimal beamforming designs under various criteria, implying that future work will leverage this insight for more complex optimization objectives. The paper also implicitly suggests that exploring other problem formulations and their equivalences could be a fruitful avenue.

Summary updated

BibTeX
@misc{yuan2013power,
  author = {Xiaojun Yuan},
  title = {{Power Minimization in Multi-pair Two-Way Relaying}},
  year = {2013},
  eprint = {1306.5513},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1306.5513},
}
Figure