Joint Path Selection and Rate Allocation Framework for 5G Self-Backhauled mmWave Networks

TL;DR

This paper presents a novel framework for joint path selection and rate allocation in 5G mmWave networks, addressing latency constraints and improving reliability through reinforcement learning and convex optimization.

Contribution

It introduces a new system design combining multiple antenna diversity, traffic splitting, and stochastic optimization for efficient path and rate management in mmWave networks.

Findings

Achieves up to 99.9999% communication reliability.

Reduces latency by over 50%.

Demonstrates a key trade-off between latency and network load.

Abstract

Owing to severe path loss and unreliable transmission over a long distance at higher frequency bands, we investigate the problem of path selection and rate allocation for multi-hop self-backhaul millimeter wave (mmWave) networks. Enabling multi-hop mmWave transmissions raises a potential issue of increased latency, and thus, in this work we aim at addressing the fundamental questions: how to select the best multi-hop paths and how to allocate rates over these paths subject to latency constraints? In this regard, we propose a new system design, which exploits multiple antenna diversity, mmWave bandwidth, and traffic splitting techniques to improve the downlink transmission. The studied problem is cast as a network utility maximization, subject to an upper delay bound constraint, network stability, and network dynamics. By leveraging stochastic optimization, the problem is decoupled into: path selection and rate allocation sub-problems, whereby a framework which selects the best paths is proposed using reinforcement learning techniques. Moreover, the rate allocation is a nonconvex program, which is converted into a convex one by using the successive convex approximation method. Via mathematical analysis, we provide a comprehensive performance analysis and convergence proofs for the proposed solution. Numerical results show that our approach ensures reliable communication with a guaranteed probability of up to $99.9999\%$, and reduces latency by $50.64\%$ and $92.9\%$ as compared to baseline models. Furthermore, the results showcase the key trade-off between latency and network arrival rate.