[Paper Review] Asymptotic Close to Optimal Resource Allocation in Centralized Multi-band Wireless Networks

Mohammadreza Darabi, Amin Roustaei|arXiv (Cornell University)|Jan 20, 2016

Advanced Wireless Network Optimization4 references3 citations

TL;DR

This paper proposes an asymptotically optimal sub-channel allocation scheme for centralized multi-band wireless networks by reformulating the non-convex sum-rate maximization problem as an assignment problem, solvable via the Hungarian algorithm. The method achieves near-optimal throughput with significantly reduced complexity compared to prior approaches, as validated by simulations showing superior performance in high-user and high-subchannel scenarios.

ABSTRACT

This paper concerns sub-channel allocation in multi-user wireless networks with a view to increasing the network throughput. It is assumed there are some sub-channels to be equally divided among active links, such that the total sum rate increases, where it is assumed each link is subject to a maximum transmit power constraint. This problem is found to be a non-convex optimization problem and is hard to deal with for large number of sub channels and/or users. However, relying on some approximation methods, it is demonstrated that the proposed sub-optimal problem has roots in combinatorial optimization, termed as Assignment problem which can be tackled through the so called Hungarian method. Simulation results demonstrate that the proposed method outperforms existing works addressed in the literature.

Motivation & Objective

• To address the challenge of maximizing sum rate in centralized multi-band wireless networks under power constraints.
• To overcome the computational intractability of the non-convex sub-channel allocation problem in large-scale networks.
• To develop a low-complexity, near-optimal solution that scales efficiently with the number of users and sub-channels.
• To demonstrate performance gains over existing sub-optimal resource allocation methods in practical multi-user scenarios.

Proposed method

• Reformulates the non-convex sum-rate maximization problem as a combinatorial assignment problem.
• Applies the Hungarian algorithm to solve the assignment problem efficiently, ensuring polynomial-time complexity.
• Uses power allocation and sub-channel assignment jointly to maximize spectral efficiency under individual transmit power constraints.
• Employs asymptotic analysis to justify the near-optimality of the solution as the number of sub-channels increases.
• Derives a sub-optimal but tractable formulation by approximating the original non-convex problem with a convex relaxation.
• Validates the method through simulation-based performance evaluation under varying user and sub-channel counts.

Experimental results

Research questions

• RQ1 How can sub-channel allocation be optimized to maximize sum rate in multi-band wireless networks with power constraints?
• RQ2 What is the computational complexity of optimal sub-channel allocation, and can it be reduced without sacrificing performance?
• RQ3 Can the non-convex sub-channel allocation problem be transformed into a solvable combinatorial optimization problem?
• RQ4 How close to optimal is the proposed solution in terms of sum rate, especially as system size increases?
• RQ5 How does the proposed method compare to existing sub-optimal algorithms in terms of throughput and scalability?

Key findings

• The proposed method achieves sum rates very close to the theoretical optimum, especially in high-dimensional scenarios with many sub-channels and users.
• The Hungarian algorithm solution provides a polynomial-time complexity solution, making it scalable for large networks.
• Simulation results confirm that the proposed method outperforms existing sub-optimal resource allocation techniques in terms of sum rate gain.
• The performance gap between the proposed solution and the optimal solution diminishes asymptotically as the number of sub-channels increases.
• The method maintains high spectral efficiency even under strict individual transmit power constraints.