Technology Category
Machine Learning: Online Learning & BanditsSearch and Optimization: Learning to SearchMultiagent Systems: Multiagent Learning
Deep Reinforcement Learning for Backhaul Link Selection for Network Slices in IAB Networks
🤖 AI Summary
To address the challenge of slice-level wireless backhaul link selection for dynamically congested base stations in 5G Integrated Access and Backhaul (IAB) networks, this paper proposes an online decision-making method based on Double Deep Q-Network (DDQN). The problem is formulated as a reinforcement learning task, where a fully connected neural network with ReLU activation maps states to actions. This work pioneers slice-granular, zero-failure, low-overhead dynamic backhaul path configuration in IAB systems—avoiding the curse of dimensionality and excessive decision latency inherent in conventional combinatorial optimization approaches. Experimental results demonstrate that the method achieves 100% test success rate within approximately 20 training episodes, significantly enhancing real-time responsiveness, adaptability, and resource efficiency of slice-based backhaul under congestion.
Technology Category
Application Category
📝 Abstract
Integrated Access and Backhaul (IAB) has been recently proposed by 3GPP to enable network operators to deploy fifth generation (5G) mobile networks with reduced costs. In this paper, we propose to use IAB to build a dynamic wireless backhaul network capable to provide additional capacity to those Base Stations (BS) experiencing congestion momentarily. As the mobile traffic demand varies across time and space, and the number of slice combinations deployed in a BS can be prohibitively high, we propose to use Deep Reinforcement Learning (DRL) to select, from a set of candidate BSs, the one that can provide backhaul capacity for each of the slices deployed in a congested BS. Our results show that a Double Deep Q-Network (DDQN) agent using a fully connected neural network and the Rectified Linear Unit (ReLU) activation function with only one hidden layer is capable to perform the BS selection task successfully, without any failure during the test phase, after being trained for around 20 episodes.
Problem
Research questions and friction points this paper is trying to address.
Optimize backhaul link selection
Enhance 5G network capacity
Apply DRL in IAB networks
Innovation
Methods, ideas, or system contributions that make the work stand out.
Deep Reinforcement Learning
Double Deep Q-Network
IAB Networks
🔎 Similar Papers
No similar papers found.
