QoS Assurance Mechanism for 5G Network Slicing Based on the Deep Reinforcement Learning PPO Algorithm

2026-05-05

AI Summary

This work addresses the challenge of providing differentiated quality-of-service (QoS) guarantees in 5G network slicing, where service heterogeneity and dynamic traffic loads complicate resource management. The authors formulate the resource allocation problem as a constrained Markov decision process that jointly optimizes bandwidth, computation, and wireless resources. They propose a novel deep reinforcement learning architecture integrating graph attention networks with bidirectional LSTM to collaboratively capture both network topology and temporal state dynamics. To enforce slice isolation while optimizing multiple QoS objectives—including latency, throughput, reliability, and fairness—the approach incorporates an adaptive Lagrangian penalty and a dynamic reward shaping mechanism. Experimental results demonstrate that the proposed method significantly outperforms existing baselines in terms of QoS satisfaction ratio, delay control, resource utilization efficiency, and convergence stability.

Abstract

With the increasing diversity of 5G service types and the intensifying dynamic fluctuations of network load, achieve differentiated quality of service assurance in a network slicing environment has become a key issue in resource management. To address this problem, this paper proposes a deep reinforcement learning mechanism for 5G network slicing quality of service assurance based on the traditional proximal policy optimization actor-critic framework. First, the slicing resource allocation is modeled as a constrained Markov decision process, jointly considering the collaborative optimization of bandwidth, computing, and wireless resources. Meanwhile, a graph attention network and bidirectional long short-term memory are introduced to extract topological correlations and temporal service features, combined with an adaptive Lagrangian penalty and dynamic reward shaping mechanism, to comprehensively optimize delay, throughput, reliability, fairness, and slice isolation performance. Experimental results show that the proposed method outperforms existing baseline models in terms of quality of service satisfaction rate, delay control, resource utilization, and convergence stability.

Problem

• 5G network slicing
• QoS assurance
• resource management
• service differentiation
• network load dynamics

Innovation

• Deep Reinforcement Learning
• Network Slicing
• Proximal Policy Optimization
• Graph Attention Network
• Resource Allocation