Spectrum-aware Multi-hop Task Routing in Vehicle-assisted Collaborative Edge Computing

This paper proposes a multi-hop task offloading framework for vehicle-assisted Multi-access Edge Computing (MEC) to enhance resource utilization and efficiency. It addresses the limitations of one-hop offloading by leveraging vehicles to form a data transportation network for multi-hop task delivery. A reinforcement learning-based approach, specifically Multi-Agent Deep Deterministic Policy Gradient (MADDPG), is developed to optimize task offloading decisions, maximizing aggregated service throughput while considering constraints on end-to-end latency, spectrum, and computing resources. Numerical results demonstrate the algorithm's superior performance compared to existing schemes.

Yiqin Deng, Haixia Zhang, Xianhao Chen, Yuguang Fang