Journal of Engineering, Project, and Production Management, 2026, 16(5), 2026-0011
A Lightweight Access Selection Algorithm for Service Types in Network Slicing Environments
1 Senior
Solutions Manager, School of Electronic Information Engineering, Geely
University, Chengdu, 641423, China, E-mail: junliuj@outlook.com
(corresponding author).
2 Director, Guangdong Vocational
College of Post and Telecom, Guangzhou 510360, China
3 Assistant to the President, Geely
University, Chengdu, Sichuan Province, China
4 Technical Lead, Sobey Video Cloud
Computing Co., Ltd, Chengdu 610000, China
5 Double Teacher, Geely
University, Chengdu, Sichuan Province, China
Engineering Management
Received April 21, 2026; revised July 1, 2026; accepted July 9th
Available online July 25, 2026
Abstract: Network slicing in 5G networks faces significant computational challenges for terminal access selection under heterogeneous service requirements. A light-weight access selection algorithm is proposed that meets the constraints and supports Quality of Service (QoS) requirements of Enhanced Mobile Broadband (eMBB), Ultra-Reliable and Low-Latency Communications (URLLC), and Massive Machine-Type Communications (mMTC) communication systems. The proposed approach uses a two-stage decision structure that combines coarse-grained filtering with fine-grained evaluation. An improved Dueling Deep Q-Network (Dueling-DQN) serves as the decision-making engine, with model compression and knowledge distillation applied to substantially reduce its size. Most importantly, adaptive weight allocation methods automatically adjust the decision variables based on network conditions at runtime. Experiments show that the algorithm maintains an overall QoS satisfaction rate above 91% while reducing computational complexity by up to 78% compared with state-of-the-art methods. The inference latency on mobile hardware is under 8ms, with power consumption reduced by up to 30%. Field deployments in Vehicle-to-Everything (V2X) and industrial IoT scenarios validate the practical effectiveness, delivering 4.2ms emergency-brake alert latency in V2X applications and managing over 1,200 concurrent devices with 96.8% data delivery reliability in manufacturing environments.
Keywords: Network slicing; resource allocation; reinforcement learning; edge computing.
Copyright © Journal of Engineering, Project, and Production Management (EPPM-Journal).
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
Requests for reprints and permissions at eppm.journal@gmail.com.
Citation: Liu, J., Wei, H., Wang, Q., Zhou, P., and Wang, T. (2026). A Lightweight Access Selection Algorithm for Service Types in Network Slicing Environments. Journal of Engineering, Project, and Production Management, 16(5), 2026-0011.
