Rethinking Maximum Flow Problem and Beamforming Design through Brain-inspired Geometric Lens

TL;DR

This paper introduces a brain-inspired geometric framework for optimizing relay positioning and beamforming in wireless networks, leveraging Riemannian metrics and machine learning to enhance data rates and network flow.

Contribution

It proposes a novel Riemannian geometry-based approach for relay placement and beamforming design, inspired by brain network classification, improving network throughput and link rates.

Findings

LEM-based relay positioning outperforms algebraic connectivity.

Geometric machine learning achieves maximum link rate quickly.

Simulation confirms effectiveness of the proposed methods.

Abstract

Increasing data rate in wireless networks can be accomplished through a two-pronged approach, which are 1) increasing the network flow rate through parallel independent routes and 2) increasing the user's link rate through beamforming codebook adaptation. Mobile relays are utilized to enable achieving these goals given their flexible positioning. First at the network level, we model regularized Laplacian matrices, which are symmetric positive definite (SPD) ones representing relay-dependent network graphs, as points over Riemannian manifolds. Inspired by the geometric classification of different tasks in the brain network, Riemannian metrics, such as Log-Euclidean metric (LEM), are utilized to choose relay positions that result in maximum LEM. Simulation results show that the proposed LEM-based relay positioning algorithm enables parallel routes and achieves maximum network flow rate, as opposed to other metrics (e.g., algebraic connectivity). Second at the link level, we design unique relay-dependent beamforming codebooks aimed to increase data rate over the spatially-correlated fading channels between a given relay and its neighboring users. To do so, we propose a geometric machine learning approach, which utilizes support vector machine (SVM) model to learn an SPD variant of the user's channel over Riemannian manifolds. Consequently, LEM-based Riemannian metric is utilized for classification of different channels, and a matched beamforming codebook is constructed accordingly. Simulation results show that the proposed geometric-based learning model achieves the maximum link rate after a short training period.