A network slicing elastic switching algorithm based on deep reinforcement learning is proposed to address the network disconnection problem caused by the mobility of VR end users, which requires fast network switching to ensure service continuity. The slicing switching problem is modeled as a Markov decision process (MDP), which matches users, slices, and base stations. Introducing Double Deep Q-network (DDQN) into software defined network (SDN) controllers to achieve intelligent network resilient switching, improving system access success rate, average rate, and service continuity. The simulation results show that the proposed algorithm has a success rate of over 85% in access, and there is also a significant improvement in user speed and service continuity.
A Network Slicing Elastic Switching Algorithm for VR Devices Based on DDQN
2024-06-24
1195036 byte
Conference paper
Electronic Resource
English
Network Slicing with Elastic SFC
IEEE | 2017
|UAV Path Planning Based on DDQN for Mountain Rescue
British Library Conference Proceedings | 2022
|Economical self-adaptive cruise control method for electric vehicle based on multi-step DDQN
European Patent Office | 2023
|