Year: 2025
Type: article
Abstract: Network slicing has emerged as a pivotal technology in addressing the diverse customization requirements of the Industrial Internet of Things (IIoT) within 5G networks, enabling the deployment of multiple logical networks over shared infrastructure. Efficient resource management in this context is essential to ensure energy efficiency and meet the stringent real-time demands of IIoT applications. This study focuses on the scheduling problem of minimizing average transmission power while maintaining Age of Information (AoI) tolerance constraints within 5G wireless network slicing. To tackle this challenge, an improved Dueling Double Deep Q-Network (D3QN) is leveraged to devise intelligent slicing schemes that dynamically allocate resources, ensuring optimal performance in time-varying wireless environments. The proposed improved D3QN approach introduces a novel heuristic-based exploration strategy that restricts action choices to the most effective options, significantly; reducing ineffective learning steps. The simulation results show that the method not only speeds up convergence considerably but also achieves lower transmit power while preserving strict AoI reliability constraints and slice isolation.
Source: Sensors
DOI: 10.3390/s25226956
Authors: Mingjiang Weng, Yixuan Bai, Xin Xie
Institution: Chongqing University of Posts and Telecommunications
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,137
Award: G7759666031
OpenAlex ID: W4416209000