DRL-Based Green Task Offloading for Content Distribution in NOMA-Enabled Cloud-Edge-End Cooperation Environments
Chao Fang, Xiangheng Meng, Zhaoming Hu, Xiaoping Yang, Fangmin Xu, Peng Li, Mianxiong Dong · 2023
With the widespread utilization of intelligent devices, massive mobile users' needs for rich multimedia services bring serious challenges in the aspects of network traffic, energy consumption and carbon emission. How to realize green content distribution by optimizing resource allocation is an urgent problem to solve in complex and dynamic networks. In this paper, we design a cross-layer cooperative scheme to promote energy efficiency in non-orthogonal multiple access (NOMA)-assisted cloud-edge-side environments. To be specific, we formulate the joint optimization issue of computation, caching and communication resources as an energy minimization model while considering request aggregation. Next, we propose a new deep reinforcement learning (DRL)-based task offloading strategy to minimize energy consumption by making optimal resource allocation decisions according to content request history and resource availability. Simulation results show that the proposed solution has better performance than current typical strategies in cloud-edge-end collaboration environments.