Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Mobility-Aware Resource Allocation for mmWave IAB Networks: A Multi-Agent Reinforcement Learning Approach

Bibo Zhang, Ilario Filippini

TL;DR

This work tackles joint flow routing and link scheduling in mobility- and blockage-affected mmWave IAB networks by deploying a scalable MARL framework. Each Tx panel is an RL agent, and centralized attention-based critics coordinate cooperation to optimize UE throughput while respecting backhaul constraints and FD/HD operation. The proposed MAAC-based method demonstrates rapid learning, outperforming baselines such as MADDPG and heuristic schemes, and remains robust across obstacle densities and user speeds. The online-training-centric design and buffer-aware rewards enable practical deployment in dynamic IAB deployments, with significant throughput gains and reasonable fairness. Overall, this approach offers a scalable, real-time solution for dense mmWave IAB networks facing mobility and stochastic blockages.

Abstract

MmWaves have been envisioned as a promising direction to provide Gbps wireless access. However, they are susceptible to high path losses and blockages, which directional antennas can only partially mitigate. That makes mmWave networks coverage-limited, thus requiring dense deployments. Integrated access and backhaul (IAB) architectures have emerged as a cost-effective solution for network densification. Resource allocation in mmWave IAB networks must face big challenges to cope with heavy temporal dynamics, such as intermittent links caused by user mobility and blockages from moving obstacles. This makes it extremely difficult to find optimal and adaptive solutions. In this article, exploiting the distributed structure of the problem, we propose a Multi-Agent Reinforcement Learning (MARL) framework to optimize user throughput via flow routing and link scheduling in mmWave IAB networks characterized by user mobility and link outages generated by moving obstacles. The proposed approach implicitly captures the environment dynamics, coordinates the interference, and manages the buffer levels of IAB relay nodes. We design different MARL components, considering full-duplex and half-duplex IAB-nodes. In addition, we provide a communication and coordination scheme for RL agents in an online training framework, addressing the feasibility issues of practical systems. Numerical results show the effectiveness of the proposed approach.

Mobility-Aware Resource Allocation for mmWave IAB Networks: A Multi-Agent Reinforcement Learning Approach

TL;DR

This work tackles joint flow routing and link scheduling in mobility- and blockage-affected mmWave IAB networks by deploying a scalable MARL framework. Each Tx panel is an RL agent, and centralized attention-based critics coordinate cooperation to optimize UE throughput while respecting backhaul constraints and FD/HD operation. The proposed MAAC-based method demonstrates rapid learning, outperforming baselines such as MADDPG and heuristic schemes, and remains robust across obstacle densities and user speeds. The online-training-centric design and buffer-aware rewards enable practical deployment in dynamic IAB deployments, with significant throughput gains and reasonable fairness. Overall, this approach offers a scalable, real-time solution for dense mmWave IAB networks facing mobility and stochastic blockages.

Abstract

MmWaves have been envisioned as a promising direction to provide Gbps wireless access. However, they are susceptible to high path losses and blockages, which directional antennas can only partially mitigate. That makes mmWave networks coverage-limited, thus requiring dense deployments. Integrated access and backhaul (IAB) architectures have emerged as a cost-effective solution for network densification. Resource allocation in mmWave IAB networks must face big challenges to cope with heavy temporal dynamics, such as intermittent links caused by user mobility and blockages from moving obstacles. This makes it extremely difficult to find optimal and adaptive solutions. In this article, exploiting the distributed structure of the problem, we propose a Multi-Agent Reinforcement Learning (MARL) framework to optimize user throughput via flow routing and link scheduling in mmWave IAB networks characterized by user mobility and link outages generated by moving obstacles. The proposed approach implicitly captures the environment dynamics, coordinates the interference, and manages the buffer levels of IAB relay nodes. We design different MARL components, considering full-duplex and half-duplex IAB-nodes. In addition, we provide a communication and coordination scheme for RL agents in an online training framework, addressing the feasibility issues of practical systems. Numerical results show the effectiveness of the proposed approach.
Paper Structure (24 sections, 14 equations, 11 figures, 3 tables, 2 algorithms)

This paper contains 24 sections, 14 equations, 11 figures, 3 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Related Work
  3. System Model
  4. Channel Model
  5. Network Designs
  6. Network Operations
  7. Adaptive Flow Routing and Link Scheduling
  8. Observation Space
  9. Action Space
  10. Reward Function
  11. Reinforcement Learning Approach
  12. Learning Framework for IAB Networks
  13. Model Deployment in IAB Networks
  14. Training Cycle Synchronization
  15. Overview of the Training Process
  16. ...and 9 more sections

Figures (11)

  • Figure 1: An example of IAB network scenario with mobile users. The dashed arrows represent user trajectories.
  • Figure 2: Blockage model considering 3D obstacles.
  • Figure 3: A top-view example of IAB network scenario and its corresponding RL formulation. (a) shows an IAB network with 1 IAB-donor and 4 IAB-nodes. Backhaul links (dashed lines) form a tree topology and the number of covered UEs is reported in each sector. A sector is shadowed if blockages have been detected in it in the previous slots. (b) details an IAB-node equipped with $N_p=4$ Tx array panels (1, 2, 3, 4), each of which manages $N_s=4$ sectors (a, b, c, d). (c) includes a table recording the binary UE presence information for the sectors shown in (a).
  • Figure 4: Basic principles of the MARL framework: (a) an overview of the MARL system architecture, (b) key components of information exchanges and model updates.
  • Figure 5: An illustrative example of system coordination solutions.
  • ...and 6 more figures